Refresh Beetle merge evaluation aggregate
Browse files- RESULTS_BEETLE_MERGE_EVAL.md +162 -117
- beetle_merge_eval_deltas.csv +0 -0
- beetle_merge_eval_long.csv +0 -0
- coverage_by_family.csv +1 -1
- coverage_by_merge.csv +107 -107
- fig_damage_vs_parent_relatedness.png +2 -2
- fig_headline_naive_vs_aligned.png +2 -2
- fig_merge_vs_joint_ceiling.png +2 -2
- summary_by_family_arm.csv +4 -4
- summary_by_family_bench_arm.csv +56 -44
- summary_by_family_benchmark.csv +23 -21
- summary_by_operator_arm.csv +11 -11
RESULTS_BEETLE_MERGE_EVAL.md
CHANGED
|
@@ -1,6 +1,6 @@
|
|
| 1 |
# Beetle merged models vs their parents — evaluation report
|
| 2 |
|
| 3 |
-
_Generated 2026-08-27
|
| 4 |
|
| 5 |
|
| 6 |
## Headline findings
|
|
@@ -8,8 +8,8 @@ _Generated 2026-08-27 02:56 UTC. **PARTIAL RUN** — see the coverage table for
|
|
| 8 |
**1. Every published `__aligned` folder is byte-identical to its `__naive` sibling.**
|
| 9 |
Verified on the LFS sha256 of `model.safetensors` for **every operator-block in all
|
| 10 |
214 merge repos that publishes both arms — 150 of 150, no exceptions** — and confirmed
|
| 11 |
-
independently by the evaluation. Of the
|
| 12 |
-
scored on the same benchmark, **
|
| 13 |
exceptions differ by one and two items respectively (|Δ| <= 2.2e-4), which is the
|
| 14 |
OOM-adaptive batch sizing changing the reduction order under bf16, not a difference
|
| 15 |
between the models. The org's own
|
|
@@ -94,8 +94,8 @@ Per-cell detail: `harness_validation_vs_beetle_analyze.csv`.
|
|
| 94 |
| arch | 1 | 6 | 6 | 1 | 0 | 0 |
|
| 95 |
| humanscale-pair | 3 | 72 | 72 | 3 | 0 | 0 |
|
| 96 |
| zoo-curriculum | 49 | 49 | 49 | 49 | 0 | 0 |
|
| 97 |
-
| zoo-language | 111 | 692 |
|
| 98 |
-
| zoo-mono_vs_joint | 50 | 50 |
|
| 99 |
|
| 100 |
Per-merge coverage: `coverage_by_merge.csv`.
|
| 101 |
|
|
@@ -103,21 +103,21 @@ Per-merge coverage: `coverage_by_merge.csv`.
|
|
| 103 |
|
| 104 |
| provenance_family | cells | merges | mean acc | mean Δ vs naive | beats naive | mean Δ vs better parent | beats better parent | mean Δ vs joint ceiling | beats ceiling |
|
| 105 |
|:--------------------|--------:|---------:|-----------:|------------------:|:--------------|--------------------------:|:----------------------|--------------------------:|:----------------|
|
| 106 |
-
| arch | 24 | 1 | 0.6736 | 0 | 0/24 | -0.
|
| 107 |
| humanscale-pair | 240 | 3 | 0.5669 | 0.0055 | 42/240 | -0.1966 | 0/240 | -0.1416 | 4/240 |
|
| 108 |
-
| zoo-curriculum | 158 | 49 | 0.
|
| 109 |
-
| zoo-language |
|
| 110 |
-
| zoo-mono_vs_joint |
|
| 111 |
|
| 112 |
### blimp_eng (chance 0.500)
|
| 113 |
|
| 114 |
| provenance_family | cells | merges | mean acc | mean Δ vs naive | beats naive | mean Δ vs better parent | beats better parent | mean Δ vs joint ceiling | beats ceiling |
|
| 115 |
|:--------------------|--------:|---------:|-----------:|------------------:|:--------------|--------------------------:|:----------------------|--------------------------:|:----------------|
|
| 116 |
-
| arch | 6 | 1 | 0.5503 | 0 | 0/6 | -0.
|
| 117 |
| humanscale-pair | 72 | 3 | 0.5242 | 0.0017 | 14/72 | -0.123 | 0/72 | -0.0345 | 1/72 |
|
| 118 |
| zoo-curriculum | 49 | 49 | 0.5925 | nan | 0/0 | -0.0137 | 5/49 | nan | 0/0 |
|
| 119 |
-
| zoo-language |
|
| 120 |
-
| zoo-mono_vs_joint |
|
| 121 |
|
| 122 |
### multiblimp:eng (chance 0.500)
|
| 123 |
|
|
@@ -125,9 +125,9 @@ Per-merge coverage: `coverage_by_merge.csv`.
|
|
| 125 |
|:--------------------|--------:|---------:|-----------:|------------------:|:--------------|--------------------------:|:----------------------|--------------------------:|:----------------|
|
| 126 |
| arch | 6 | 1 | 0.7147 | 0 | 0/6 | -0.0152 | 0/6 | nan | 0/0 |
|
| 127 |
| humanscale-pair | 72 | 3 | 0.5609 | 0.0093 | 13/72 | -0.2287 | 0/72 | -0.1512 | 0/72 |
|
| 128 |
-
| zoo-curriculum | 49 | 49 | 0.
|
| 129 |
-
| zoo-language |
|
| 130 |
-
| zoo-mono_vs_joint |
|
| 131 |
|
| 132 |
## Worked example — the three named `beetle-humanscale-*-eng` pairs
|
| 133 |
|
|
@@ -160,27 +160,71 @@ shown against both bars so the loss to the better parent stays visible):
|
|
| 160 |
## The gradient that actually explains the results
|
| 161 |
|
| 162 |
Damage from merging is not a constant of the method; it tracks **how related the two
|
| 163 |
-
parents are**
|
| 164 |
-
|
| 165 |
-
|
| 166 |
-
|
| 167 |
-
|
| 168 |
-
|
| 169 |
-
|
| 170 |
-
|
| 171 |
-
|
| 172 |
-
|
| 173 |
-
|
| 174 |
-
|
| 175 |
-
|
| 176 |
-
|
| 177 |
-
|
| 178 |
-
|
| 179 |
-
|
| 180 |
-
|
| 181 |
-
|
| 182 |
-
|
| 183 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 184 |
|
| 185 |
## The flattering comparison, stated honestly
|
| 186 |
|
|
@@ -192,9 +236,9 @@ better merge. Every cell where both bars are computable:
|
|
| 192 |
|:--------------------|-----------------------:|--------------:|-----------------------------------------:|------------------:|
|
| 193 |
| arch | 24 | 0 | 0 | 0 |
|
| 194 |
| humanscale-pair | 240 | 42 | 42 | 0 |
|
| 195 |
-
| zoo-language |
|
| 196 |
|
| 197 |
-
**
|
| 198 |
Reporting Δ vs naive alone would present those as successes.
|
| 199 |
|
| 200 |
## Honest verdict per family
|
|
@@ -205,7 +249,7 @@ Scored **6 of 6** published rungs across 1 of 1 repos (24 model-benchmark cells)
|
|
| 205 |
|
| 206 |
A single cross-architecture pair (pico x MoE).
|
| 207 |
|
| 208 |
-
- **vs better parent**:
|
| 209 |
- **vs joint ceiling**: no jointly-trained counterpart published for this family.
|
| 210 |
- **vs naive**: 0/24 cells above; mean Δ +0.0000.
|
| 211 |
- Mean accuracy 0.6736 against a chance floor of 0.500; 0% of cells fall within 2 points of chance.
|
|
@@ -227,32 +271,32 @@ Scored **49 of 49** published rungs across 49 of 49 repos (158 model-benchmark c
|
|
| 227 |
|
| 228 |
Published rungs are `qmd_guided__aligned` only — one rung per repo, no naive and no transport sibling. Δ vs naive is unavailable, and since every aligned folder elsewhere in the org is byte-identical to its naive sibling, these are best read as naive merges.
|
| 229 |
|
| 230 |
-
- **vs better parent**:
|
| 231 |
- **vs joint ceiling**: no jointly-trained counterpart published for this family.
|
| 232 |
- **vs naive**: **no naive artefact exists** for these repos, so this bar cannot be evaluated. Not a failure — an absence.
|
| 233 |
-
- Mean accuracy 0.
|
| 234 |
|
| 235 |
### zoo-language
|
| 236 |
|
| 237 |
-
Scored **
|
| 238 |
|
| 239 |
Largest family. Carries ties/average/task_arithmetic in naive, aligned and transport arms, so it is the only family with broad arm coverage — but the aligned arm is byte-identical to naive throughout.
|
| 240 |
|
| 241 |
-
- **vs better parent**:
|
| 242 |
- **vs joint ceiling**: no jointly-trained counterpart published for this family.
|
| 243 |
-
- **vs naive**:
|
| 244 |
-
- Mean accuracy 0.
|
| 245 |
|
| 246 |
### zoo-mono_vs_joint
|
| 247 |
|
| 248 |
-
Scored **
|
| 249 |
|
| 250 |
Parent B *is* the jointly-trained model, so the joint ceiling is built in and `d_ceiling` is meaningful here. Published rungs are `qmd_guided__aligned` and `qmd_guided__transport` only — no naive folder exists, so Δ vs naive is unavailable by construction.
|
| 251 |
|
| 252 |
-
- **vs better parent**:
|
| 253 |
-
- **vs joint ceiling**:
|
| 254 |
- **vs naive**: **no naive artefact exists** for these repos, so this bar cannot be evaluated. Not a failure — an absence.
|
| 255 |
-
- Mean accuracy 0.
|
| 256 |
|
| 257 |
## Which merge operator does least damage
|
| 258 |
|
|
@@ -341,11 +385,11 @@ a property of the benchmark, not a gap in this run.
|
|
| 341 |
This is the question only `zoo-mono_vs_joint` and `humanscale-pair` can answer, because
|
| 342 |
they are the two families with a published jointly-trained counterpart.
|
| 343 |
|
| 344 |
-
- Across **
|
| 345 |
-
model in **
|
| 346 |
-
- Per family: humanscale-pair 4/240, zoo-mono_vs_joint
|
| 347 |
|
| 348 |
-
**A caveat that changes how the number should be read.** In **
|
| 349 |
jointly-trained model is itself *below* the better of the two parents — bilingual joint
|
| 350 |
training costs English BLiMP relative to a monolingual English parent, for instance
|
| 351 |
(0.573 for the nld-eng joint model vs 0.647 for monolingual English). So "ceiling" here
|
|
@@ -363,67 +407,68 @@ See `fig_merge_vs_joint_ceiling.png`.
|
|
| 363 |
|
| 364 |
## Reference models (parents and joint ceilings)
|
| 365 |
|
| 366 |
-
| merge_id
|
| 367 |
-
|:------------------------------------------------------------------------------------|------------:|-----------:|---------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------:|
|
| 368 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-deu-eng
|
| 369 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-eus-eng
|
| 370 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-fil-eng
|
| 371 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-hin-eng
|
| 372 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-ita-eng
|
| 373 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-nld-eng
|
| 374 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-pol-eng
|
| 375 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-rus-eng
|
| 376 |
-
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-tur-eng
|
| 377 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-classroom-20-b4-fineweb-100m-nld-eng
|
| 378 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-deu-eng
|
| 379 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-eus-eng
|
| 380 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-fil-eng
|
| 381 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-hin-eng
|
| 382 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-ita-eng
|
| 383 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-nld-eng
|
| 384 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-pol-eng
|
| 385 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-rus-eng
|
| 386 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-tur-eng
|
| 387 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-deu-eng
|
| 388 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-eus-eng
|
| 389 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-fil-eng
|
| 390 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-hin-eng
|
| 391 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-ita-eng
|
| 392 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-nld-eng
|
| 393 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-pol-eng
|
| 394 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-rus-eng
|
| 395 |
-
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-tur-eng
|
| 396 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-deu
|
| 397 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-ell
|
| 398 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-est
|
| 399 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-fin
|
| 400 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-hin
|
| 401 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-isl
|
| 402 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-ita
|
| 403 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-jpn
|
| 404 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-kor
|
| 405 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-nld
|
| 406 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-pol
|
| 407 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-rus
|
| 408 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-spa
|
| 409 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-tur
|
| 410 |
-
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-zho
|
| 411 |
-
| Beetle-FineWeb-2B/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-2b-nld-eng
|
| 412 |
-
| Beetle-HumanScale/beetle-bilingual-balanced-b1-humanscale-deu-eng
|
| 413 |
-
| Beetle-HumanScale/beetle-bilingual-balanced-b1-humanscale-nld-eng
|
| 414 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-classroom-20-b4-humanscale-deu-eng
|
| 415 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-classroom-20-b4-humanscale-nld-eng
|
| 416 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-0-100-b3-humanscale-nld-eng
|
| 417 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-33-67-b3-humanscale-deu-eng
|
| 418 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-33-67-b3-humanscale-nld-eng
|
| 419 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-33-67-b3-humanscale-zho-eng
|
| 420 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-simultaneous-b2-humanscale-deu-eng
|
| 421 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-simultaneous-b2-humanscale-nld-eng
|
| 422 |
-
| Beetle-HumanScale/beetle-bilingual-l2-50-simultaneous-b2-humanscale-zho-eng
|
| 423 |
-
| Beetle-HumanScale/beetle-bilingual-l2-80-late-b5-humanscale-deu-eng
|
| 424 |
-
| Beetle-HumanScale/beetle-bilingual-l2-80-late-b5-humanscale-nld-eng
|
| 425 |
-
| Beetle-HumanScale/beetle-bilingual-l2-80-late-b5-humanscale-zho-eng
|
| 426 |
-
| Beetle-HumanScale/beetle-monolingual-humanscale-deu
|
| 427 |
-
| Beetle-HumanScale/beetle-monolingual-humanscale-eng
|
| 428 |
-
| Beetle-HumanScale/beetle-monolingual-humanscale-nld
|
| 429 |
-
| Beetle-HumanScale/beetle-monolingual-humanscale-zho
|
|
|
|
|
|
| 1 |
# Beetle merged models vs their parents — evaluation report
|
| 2 |
|
| 3 |
+
_Generated 2026-08-27 04:22 UTC. **PARTIAL RUN** — see the coverage table for exactly what is and is not scored._
|
| 4 |
|
| 5 |
|
| 6 |
## Headline findings
|
|
|
|
| 8 |
**1. Every published `__aligned` folder is byte-identical to its `__naive` sibling.**
|
| 9 |
Verified on the LFS sha256 of `model.safetensors` for **every operator-block in all
|
| 10 |
214 merge repos that publishes both arms — 150 of 150, no exceptions** — and confirmed
|
| 11 |
+
independently by the evaluation. Of the 543 aligned rungs that have a naive sibling
|
| 12 |
+
scored on the same benchmark, **534 give Δ vs naive of exactly 0.0000**; the 9
|
| 13 |
exceptions differ by one and two items respectively (|Δ| <= 2.2e-4), which is the
|
| 14 |
OOM-adaptive batch sizing changing the reduction order under bf16, not a difference
|
| 15 |
between the models. The org's own
|
|
|
|
| 94 |
| arch | 1 | 6 | 6 | 1 | 0 | 0 |
|
| 95 |
| humanscale-pair | 3 | 72 | 72 | 3 | 0 | 0 |
|
| 96 |
| zoo-curriculum | 49 | 49 | 49 | 49 | 0 | 0 |
|
| 97 |
+
| zoo-language | 111 | 692 | 692 | 111 | 0 | 0 |
|
| 98 |
+
| zoo-mono_vs_joint | 50 | 50 | 50 | 50 | 0 | 0 |
|
| 99 |
|
| 100 |
Per-merge coverage: `coverage_by_merge.csv`.
|
| 101 |
|
|
|
|
| 103 |
|
| 104 |
| provenance_family | cells | merges | mean acc | mean Δ vs naive | beats naive | mean Δ vs better parent | beats better parent | mean Δ vs joint ceiling | beats ceiling |
|
| 105 |
|:--------------------|--------:|---------:|-----------:|------------------:|:--------------|--------------------------:|:----------------------|--------------------------:|:----------------|
|
| 106 |
+
| arch | 24 | 1 | 0.6736 | 0 | 0/24 | -0.0656 | 0/24 | nan | 0/0 |
|
| 107 |
| humanscale-pair | 240 | 3 | 0.5669 | 0.0055 | 42/240 | -0.1966 | 0/240 | -0.1416 | 4/240 |
|
| 108 |
+
| zoo-curriculum | 158 | 49 | 0.739 | nan | 0/0 | -0.0184 | 16/158 | nan | 0/0 |
|
| 109 |
+
| zoo-language | 2521 | 111 | 0.6079 | 0.0019 | 266/1367 | -0.1023 | 332/2521 | nan | 0/0 |
|
| 110 |
+
| zoo-mono_vs_joint | 166 | 50 | 0.6582 | nan | 0/0 | -0.0839 | 11/166 | -0.0712 | 36/166 |
|
| 111 |
|
| 112 |
### blimp_eng (chance 0.500)
|
| 113 |
|
| 114 |
| provenance_family | cells | merges | mean acc | mean Δ vs naive | beats naive | mean Δ vs better parent | beats better parent | mean Δ vs joint ceiling | beats ceiling |
|
| 115 |
|:--------------------|--------:|---------:|-----------:|------------------:|:--------------|--------------------------:|:----------------------|--------------------------:|:----------------|
|
| 116 |
+
| arch | 6 | 1 | 0.5503 | 0 | 0/6 | -0.0089 | 0/6 | nan | 0/0 |
|
| 117 |
| humanscale-pair | 72 | 3 | 0.5242 | 0.0017 | 14/72 | -0.123 | 0/72 | -0.0345 | 1/72 |
|
| 118 |
| zoo-curriculum | 49 | 49 | 0.5925 | nan | 0/0 | -0.0137 | 5/49 | nan | 0/0 |
|
| 119 |
+
| zoo-language | 692 | 111 | 0.5249 | 0.0003 | 69/373 | -0.0166 | 146/692 | nan | 0/0 |
|
| 120 |
+
| zoo-mono_vs_joint | 50 | 50 | 0.5602 | nan | 0/0 | -0.0368 | 6/50 | -0.0166 | 14/50 |
|
| 121 |
|
| 122 |
### multiblimp:eng (chance 0.500)
|
| 123 |
|
|
|
|
| 125 |
|:--------------------|--------:|---------:|-----------:|------------------:|:--------------|--------------------------:|:----------------------|--------------------------:|:----------------|
|
| 126 |
| arch | 6 | 1 | 0.7147 | 0 | 0/6 | -0.0152 | 0/6 | nan | 0/0 |
|
| 127 |
| humanscale-pair | 72 | 3 | 0.5609 | 0.0093 | 13/72 | -0.2287 | 0/72 | -0.1512 | 0/72 |
|
| 128 |
+
| zoo-curriculum | 49 | 49 | 0.7711 | nan | 0/0 | -0.0194 | 4/49 | nan | 0/0 |
|
| 129 |
+
| zoo-language | 692 | 111 | 0.6656 | 0.0013 | 66/373 | -0.0212 | 186/692 | nan | 0/0 |
|
| 130 |
+
| zoo-mono_vs_joint | 50 | 50 | 0.6825 | nan | 0/0 | -0.0778 | 3/50 | -0.0609 | 12/50 |
|
| 131 |
|
| 132 |
## Worked example — the three named `beetle-humanscale-*-eng` pairs
|
| 133 |
|
|
|
|
| 160 |
## The gradient that actually explains the results
|
| 161 |
|
| 162 |
Damage from merging is not a constant of the method; it tracks **how related the two
|
| 163 |
+
parents are**. Two controls are needed before that gradient can be read, and both matter:
|
| 164 |
+
|
| 165 |
+
**Control 1 — a vacuous bar.** In **717 of 3109** cells the better parent is itself within
|
| 166 |
+
10 points of chance: a `deu`+`ell` pair scored on *English* BLiMP has no parent that can
|
| 167 |
+
do the task, so "the merge nearly matches its better parent" is true and meaningless.
|
| 168 |
+
Those cells are excluded below (`parent_competent` in the delta CSV keeps them
|
| 169 |
+
inspectable). Without this exclusion `zoo-language` looks far gentler than it is
|
| 170 |
+
(-0.0173 on English BLiMP, purely because both parents sit at chance there).
|
| 171 |
+
|
| 172 |
+
**Control 2 — operator mix.** The families were not given the same operators:
|
| 173 |
+
`humanscale-pair` carries all eight including the destructive `dare`/`csp`/`svc`/`opcm`,
|
| 174 |
+
while `zoo-language` is mostly `ties`/`average`/`task_arithmetic`/`qmd_guided`. Comparing
|
| 175 |
+
families without controlling for that compares operator sets, not parent relatedness.
|
| 176 |
+
|
| 177 |
+
### Competent-parent cells, all operators
|
| 178 |
+
|
| 179 |
+
| provenance_family | parents are | cells | mean_better_parent | mean_d_parent | median_d_parent | beats_parent |
|
| 180 |
+
|:--------------------|:-----------------------------------------------------------|--------:|---------------------:|----------------:|------------------:|---------------:|
|
| 181 |
+
| zoo-curriculum | two curriculum variants of the SAME bilingual model | 136 | 0.7888 | -0.0201 | -0.0208 | 12 |
|
| 182 |
+
| arch | the same pair in two different architectures (pico vs MoE) | 18 | 0.7991 | -0.0845 | -0.0622 | 0 |
|
| 183 |
+
| zoo-mono_vs_joint | a monolingual and a bilingual sharing one of its languages | 139 | 0.7784 | -0.0943 | -0.061 | 10 |
|
| 184 |
+
| zoo-language | two separately pretrained monolinguals | 1859 | 0.772 | -0.1342 | -0.1137 | 188 |
|
| 185 |
+
| humanscale-pair | two separately pretrained monolinguals | 240 | 0.7635 | -0.1966 | -0.1565 | 0 |
|
| 186 |
+
|
| 187 |
+
### Competent-parent cells, restricted to the operators the families share (average, ties, task_arithmetic)
|
| 188 |
+
|
| 189 |
+
| provenance_family | parents are | cells | mean_better_parent | mean_d_parent | median_d_parent | beats_parent |
|
| 190 |
+
|:--------------------|:-----------------------------------------------------------|--------:|---------------------:|----------------:|------------------:|---------------:|
|
| 191 |
+
| arch | the same pair in two different architectures (pico vs MoE) | 18 | 0.7991 | -0.0845 | -0.0622 | 0 |
|
| 192 |
+
| humanscale-pair | two separately pretrained monolinguals | 90 | 0.7635 | -0.1375 | -0.1283 | 0 |
|
| 193 |
+
| zoo-language | two separately pretrained monolinguals | 1519 | 0.772 | -0.139 | -0.1247 | 155 |
|
| 194 |
+
|
| 195 |
+
**The gradient survives both controls, and the second table is the real check.**
|
| 196 |
+
`humanscale-pair` and `zoo-language` carry the *same* description — two separately
|
| 197 |
+
pretrained monolinguals — and once the vacuous cells and the operator mix are controlled
|
| 198 |
+
they collapse onto essentially the same value (-0.1390 vs -0.1375). That is what should happen if the driver really is
|
| 199 |
+
parent relatedness rather than family identity, and it is the strongest evidence here
|
| 200 |
+
that the ordering is not an artefact of how the sweep was constructed.
|
| 201 |
+
|
| 202 |
+
So, on competent-parent cells: merging two curriculum variants of one bilingual model
|
| 203 |
+
costs about **2.0 accuracy points** against the better parent, while merging two
|
| 204 |
+
separately pretrained monolinguals costs **13.8-13.9** on the shared operators and
|
| 205 |
+
up to **19.7** with all eight in play (the extra four are the destructive ones).
|
| 206 |
+
|
| 207 |
+
One honest limit on the like-for-like table: `zoo-curriculum` and `zoo-mono_vs_joint`
|
| 208 |
+
publish **only** `qmd_guided`, so they cannot appear in a shared-operator comparison at
|
| 209 |
+
all — the operator control can be applied to the two monolingual-pair families against
|
| 210 |
+
each other, which is the comparison that matters for the confound, but not across the
|
| 211 |
+
whole gradient. The curriculum figure above is therefore its `qmd_guided` value, and the
|
| 212 |
+
gradient's two endpoints are not fully operator-matched. What *is* operator-matched is
|
| 213 |
+
the pair that shares a description, and they agree.
|
| 214 |
+
|
| 215 |
+
*Curriculum siblings are mergeable;
|
| 216 |
+
cross-lingual pretrained pairs are not.* A pooled "does merging work" number is not
|
| 217 |
+
meaningful for this collection; the per-family tables are the unit of interpretation.
|
| 218 |
+
|
| 219 |
+
For reference, the uncontrolled version of the same table (both confounds left in):
|
| 220 |
+
|
| 221 |
+
| provenance_family | parents are | cells | mean_better_parent | mean_d_parent | median_d_parent | beats_parent |
|
| 222 |
+
|:--------------------|:-----------------------------------------------------------|--------:|---------------------:|----------------:|------------------:|---------------:|
|
| 223 |
+
| zoo-curriculum | two curriculum variants of the SAME bilingual model | 158 | 0.7573 | -0.0184 | -0.0193 | 16 |
|
| 224 |
+
| arch | the same pair in two different architectures (pico vs MoE) | 24 | 0.7392 | -0.0656 | -0.0377 | 0 |
|
| 225 |
+
| zoo-mono_vs_joint | a monolingual and a bilingual sharing one of its languages | 166 | 0.7422 | -0.0839 | -0.0448 | 11 |
|
| 226 |
+
| zoo-language | two separately pretrained monolinguals | 2521 | 0.7102 | -0.1023 | -0.0429 | 332 |
|
| 227 |
+
| humanscale-pair | two separately pretrained monolinguals | 240 | 0.7635 | -0.1966 | -0.1565 | 0 |
|
| 228 |
|
| 229 |
## The flattering comparison, stated honestly
|
| 230 |
|
|
|
|
| 236 |
|:--------------------|-----------------------:|--------------:|-----------------------------------------:|------------------:|
|
| 237 |
| arch | 24 | 0 | 0 | 0 |
|
| 238 |
| humanscale-pair | 240 | 42 | 42 | 0 |
|
| 239 |
+
| zoo-language | 1367 | 266 | 223 | 43 |
|
| 240 |
|
| 241 |
+
**265 of 308 cells that beat naive still lose to the better parent.**
|
| 242 |
Reporting Δ vs naive alone would present those as successes.
|
| 243 |
|
| 244 |
## Honest verdict per family
|
|
|
|
| 249 |
|
| 250 |
A single cross-architecture pair (pico x MoE).
|
| 251 |
|
| 252 |
+
- **vs better parent**: 0/24 cells above; mean Δ -0.0656.
|
| 253 |
- **vs joint ceiling**: no jointly-trained counterpart published for this family.
|
| 254 |
- **vs naive**: 0/24 cells above; mean Δ +0.0000.
|
| 255 |
- Mean accuracy 0.6736 against a chance floor of 0.500; 0% of cells fall within 2 points of chance.
|
|
|
|
| 271 |
|
| 272 |
Published rungs are `qmd_guided__aligned` only — one rung per repo, no naive and no transport sibling. Δ vs naive is unavailable, and since every aligned folder elsewhere in the org is byte-identical to its naive sibling, these are best read as naive merges.
|
| 273 |
|
| 274 |
+
- **vs better parent**: 16/158 cells above; mean Δ -0.0184.
|
| 275 |
- **vs joint ceiling**: no jointly-trained counterpart published for this family.
|
| 276 |
- **vs naive**: **no naive artefact exists** for these repos, so this bar cannot be evaluated. Not a failure — an absence.
|
| 277 |
+
- Mean accuracy 0.7390 against a chance floor of 0.500; 0% of cells fall within 2 points of chance.
|
| 278 |
|
| 279 |
### zoo-language
|
| 280 |
|
| 281 |
+
Scored **692 of 692** published rungs across 111 of 111 repos (2521 model-benchmark cells).
|
| 282 |
|
| 283 |
Largest family. Carries ties/average/task_arithmetic in naive, aligned and transport arms, so it is the only family with broad arm coverage — but the aligned arm is byte-identical to naive throughout.
|
| 284 |
|
| 285 |
+
- **vs better parent**: 332/2521 cells above; mean Δ -0.1023.
|
| 286 |
- **vs joint ceiling**: no jointly-trained counterpart published for this family.
|
| 287 |
+
- **vs naive**: 266/1367 cells above; mean Δ +0.0019.
|
| 288 |
+
- Mean accuracy 0.6079 against a chance floor of 0.500; 16% of cells fall within 2 points of chance.
|
| 289 |
|
| 290 |
### zoo-mono_vs_joint
|
| 291 |
|
| 292 |
+
Scored **50 of 50** published rungs across 50 of 50 repos (166 model-benchmark cells).
|
| 293 |
|
| 294 |
Parent B *is* the jointly-trained model, so the joint ceiling is built in and `d_ceiling` is meaningful here. Published rungs are `qmd_guided__aligned` and `qmd_guided__transport` only — no naive folder exists, so Δ vs naive is unavailable by construction.
|
| 295 |
|
| 296 |
+
- **vs better parent**: 11/166 cells above; mean Δ -0.0839.
|
| 297 |
+
- **vs joint ceiling**: 36/166 cells above; mean Δ -0.0712.
|
| 298 |
- **vs naive**: **no naive artefact exists** for these repos, so this bar cannot be evaluated. Not a failure — an absence.
|
| 299 |
+
- Mean accuracy 0.6582 against a chance floor of 0.500; 10% of cells fall within 2 points of chance.
|
| 300 |
|
| 301 |
## Which merge operator does least damage
|
| 302 |
|
|
|
|
| 385 |
This is the question only `zoo-mono_vs_joint` and `humanscale-pair` can answer, because
|
| 386 |
they are the two families with a published jointly-trained counterpart.
|
| 387 |
|
| 388 |
+
- Across **406** cells with a ceiling, the merge is at or above the jointly-trained
|
| 389 |
+
model in **40**; mean gap **-0.1128** accuracy points.
|
| 390 |
+
- Per family: humanscale-pair 4/240, zoo-mono_vs_joint 36/166.
|
| 391 |
|
| 392 |
+
**A caveat that changes how the number should be read.** In **382 of 406** cells the
|
| 393 |
jointly-trained model is itself *below* the better of the two parents — bilingual joint
|
| 394 |
training costs English BLiMP relative to a monolingual English parent, for instance
|
| 395 |
(0.573 for the nld-eng joint model vs 0.647 for monolingual English). So "ceiling" here
|
|
|
|
| 407 |
|
| 408 |
## Reference models (parents and joint ceilings)
|
| 409 |
|
| 410 |
+
| merge_id | blimp_eng | blimp_nl | bliss | multiblimp:deu | multiblimp:ell | multiblimp:eng | multiblimp:est | multiblimp:eus | multiblimp:fin | multiblimp:hin | multiblimp:isl | multiblimp:ita | multiblimp:nld | multiblimp:pol | multiblimp:rus | multiblimp:spa | multiblimp:tur | zhoblimp |
|
| 411 |
+
|:-----------------------------------------------------------------------------------------------|------------:|-----------:|---------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------------:|-----------:|
|
| 412 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-deu-eng | 0.636 | nan | nan | 0.8864 | nan | 0.8766 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 413 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-eus-eng | 0.6208 | nan | nan | nan | nan | 0.8545 | nan | 0.9744 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 414 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-fil-eng | 0.6496 | nan | nan | nan | nan | 0.8597 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 415 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-hin-eng | 0.6382 | nan | nan | nan | nan | 0.8636 | nan | nan | nan | 0.8756 | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 416 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-ita-eng | 0.6274 | nan | nan | nan | nan | 0.8779 | nan | nan | nan | nan | nan | 0.8169 | nan | nan | nan | nan | nan | nan |
|
| 417 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-nld-eng | 0.6386 | 0.7452 | nan | nan | nan | 0.8805 | nan | nan | nan | nan | nan | nan | 0.8837 | nan | nan | nan | nan | nan |
|
| 418 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-pol-eng | 0.6222 | nan | nan | nan | nan | 0.8662 | nan | nan | nan | nan | nan | nan | nan | 0.7775 | nan | nan | nan | nan |
|
| 419 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-rus-eng | 0.6089 | nan | nan | nan | nan | 0.8506 | nan | nan | nan | nan | nan | nan | nan | nan | 0.7743 | nan | nan | nan |
|
| 420 |
+
| Beetle-FineWeb-100M/beetle-bilingual-balanced-b1-fineweb-100m-tur-eng | 0.6113 | nan | nan | nan | nan | 0.8442 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | 0.7646 | nan |
|
| 421 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-classroom-20-b4-fineweb-100m-nld-eng | 0.5836 | 0.7577 | nan | nan | nan | 0.7364 | nan | nan | nan | nan | nan | nan | 0.9163 | nan | nan | nan | nan | nan |
|
| 422 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-deu-eng | 0.5479 | nan | nan | 0.9169 | nan | 0.7208 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 423 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-eus-eng | 0.5403 | nan | nan | nan | nan | 0.7039 | nan | 0.9744 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 424 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-fil-eng | 0.5857 | nan | nan | nan | nan | 0.7571 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 425 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-hin-eng | 0.5498 | nan | nan | nan | nan | 0.6831 | nan | nan | nan | 0.9178 | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 426 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-ita-eng | 0.5426 | nan | nan | nan | nan | 0.7273 | nan | nan | nan | nan | nan | 0.8486 | nan | nan | nan | nan | nan | nan |
|
| 427 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-nld-eng | 0.5738 | 0.7648 | nan | nan | nan | 0.7351 | nan | nan | nan | nan | nan | nan | 0.9271 | nan | nan | nan | nan | nan |
|
| 428 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-pol-eng | 0.5477 | nan | nan | nan | nan | 0.6909 | nan | nan | nan | nan | nan | nan | nan | 0.8139 | nan | nan | nan | nan |
|
| 429 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-rus-eng | 0.5336 | nan | nan | nan | nan | 0.6987 | nan | nan | nan | nan | nan | nan | nan | nan | 0.8207 | nan | nan | nan |
|
| 430 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-100m-tur-eng | 0.5333 | nan | nan | nan | nan | 0.6974 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | 0.806 | nan |
|
| 431 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-deu-eng | 0.5282 | nan | nan | 0.9208 | nan | 0.7117 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 432 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-eus-eng | 0.5264 | nan | nan | nan | nan | 0.6987 | nan | 0.967 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 433 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-fil-eng | 0.5902 | nan | nan | nan | nan | 0.7545 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 434 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-hin-eng | 0.5457 | nan | nan | nan | nan | 0.6714 | nan | nan | nan | 0.9164 | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 435 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-ita-eng | 0.5373 | nan | nan | nan | nan | 0.7013 | nan | nan | nan | nan | nan | 0.8483 | nan | nan | nan | nan | nan | nan |
|
| 436 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-nld-eng | 0.5761 | 0.7609 | nan | nan | nan | 0.7377 | nan | nan | nan | nan | nan | nan | 0.9112 | nan | nan | nan | nan | nan |
|
| 437 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-pol-eng | 0.5342 | nan | nan | nan | nan | 0.7078 | nan | nan | nan | nan | nan | nan | nan | 0.8243 | nan | nan | nan | nan |
|
| 438 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-rus-eng | 0.5369 | nan | nan | nan | nan | 0.6896 | nan | nan | nan | nan | nan | nan | nan | nan | 0.8129 | nan | nan | nan |
|
| 439 |
+
| Beetle-FineWeb-100M/beetle-bilingual-l2-50-simultaneous-b2-fineweb-100m-tur-eng | 0.5327 | nan | nan | nan | nan | 0.7117 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | 0.8106 | nan |
|
| 440 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-deu | 0.5235 | 0.5009 | nan | 0.8925 | 0.6369 | 0.6701 | nan | nan | 0.5696 | 0.566 | 0.6715 | 0.5175 | 0.4672 | 0.5966 | 0.5391 | 0.6348 | 0.6665 | 0.4735 |
|
| 441 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-ell | 0.528 | 0.4781 | nan | 0.631 | 0.9033 | 0.6753 | nan | nan | 0.5607 | 0.4858 | 0.658 | 0.4668 | 0.4758 | 0.607 | 0.4995 | 0.6596 | 0.6349 | 0.5293 |
|
| 442 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-est | 0.5178 | nan | nan | nan | nan | 0.6805 | 0.8012 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 443 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-fin | 0.5044 | 0.5088 | nan | 0.6671 | 0.5693 | 0.6896 | nan | nan | 0.8389 | 0.5252 | 0.6491 | 0.5092 | 0.4603 | 0.6039 | 0.5483 | 0.7009 | 0.6544 | 0.5047 |
|
| 444 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-hin | 0.5308 | 0.4632 | nan | 0.5688 | 0.5757 | 0.6766 | nan | nan | 0.565 | 0.8742 | 0.6498 | 0.4405 | 0.4612 | 0.6103 | 0.517 | 0.6517 | 0.6567 | 0.5116 |
|
| 445 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-isl | 0.5096 | 0.4716 | nan | 0.6406 | 0.5648 | 0.6701 | nan | nan | 0.5654 | 0.5487 | 0.9204 | 0.5232 | 0.4856 | 0.5905 | 0.5222 | 0.6765 | 0.6854 | 0.5105 |
|
| 446 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-ita | 0.5344 | 0.4691 | nan | 0.6619 | 0.5593 | 0.6922 | nan | nan | 0.5521 | 0.5128 | 0.6419 | 0.8089 | 0.574 | 0.6027 | 0.5459 | 0.697 | 0.6711 | 0.5231 |
|
| 447 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-jpn | 0.5374 | 0.4657 | nan | 0.6214 | 0.552 | 0.6558 | nan | nan | 0.5872 | 0.5294 | 0.6476 | 0.5092 | 0.5272 | 0.6085 | 0.5577 | 0.6793 | 0.6206 | 0.521 |
|
| 448 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-kor | 0.5207 | 0.4991 | nan | 0.6449 | 0.5894 | 0.6442 | nan | nan | 0.5342 | 0.5314 | 0.6694 | 0.4872 | 0.5641 | 0.6149 | 0.5371 | 0.6277 | 0.6315 | 0.4971 |
|
| 449 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-nld | 0.5527 | 0.7346 | nan | 0.6867 | 0.5803 | 0.6779 | nan | nan | 0.5755 | 0.5543 | 0.6566 | 0.5275 | 0.8812 | 0.6143 | 0.494 | 0.6974 | 0.659 | 0.4956 |
|
| 450 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-pol | 0.5303 | 0.5149 | nan | 0.6371 | 0.6487 | 0.6558 | nan | nan | 0.5739 | 0.5508 | 0.6583 | 0.5405 | 0.4805 | 0.7882 | 0.535 | 0.6781 | 0.6418 | 0.4689 |
|
| 451 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-rus | 0.5479 | 0.4768 | nan | 0.6079 | 0.6122 | 0.6545 | nan | nan | 0.5728 | 0.4824 | 0.6598 | 0.4538 | 0.4758 | 0.6042 | 0.7941 | 0.6907 | 0.6487 | 0.5077 |
|
| 452 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-spa | 0.5495 | 0.4896 | nan | 0.6249 | 0.5931 | 0.7182 | nan | nan | 0.5802 | 0.5128 | 0.6483 | 0.5328 | 0.4895 | 0.6103 | 0.5235 | 0.8977 | 0.6561 | 0.4826 |
|
| 453 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-tur | 0.522 | 0.4717 | nan | 0.6397 | 0.5383 | 0.6519 | nan | nan | 0.5794 | 0.5218 | 0.6348 | 0.5258 | 0.5307 | 0.6186 | 0.5042 | 0.7135 | 0.7664 | 0.4747 |
|
| 454 |
+
| Beetle-FineWeb-100M/beetle-monolingual-fineweb-100m-zho | 0.5303 | 0.4631 | nan | 0.5692 | 0.5347 | 0.6844 | nan | nan | 0.5658 | 0.5135 | 0.6516 | 0.4968 | 0.474 | 0.6027 | 0.5188 | 0.6812 | 0.6688 | 0.6458 |
|
| 455 |
+
| Beetle-FineWeb-2B/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-2b-nld-eng | 0.5552 | 0.7683 | nan | nan | nan | 0.7299 | nan | nan | nan | nan | nan | nan | 0.8992 | nan | nan | nan | nan | nan |
|
| 456 |
+
| Beetle-HumanScale/beetle-bilingual-balanced-b1-humanscale-deu-eng | 0.6546 | nan | nan | 0.869 | nan | 0.8052 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 457 |
+
| Beetle-HumanScale/beetle-bilingual-balanced-b1-humanscale-nld-eng | 0.6906 | 0.7687 | nan | nan | nan | 0.8078 | nan | nan | nan | nan | nan | nan | 0.8752 | nan | nan | nan | nan | nan |
|
| 458 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-classroom-20-b4-humanscale-deu-eng | 0.5702 | nan | nan | 0.8977 | nan | 0.7221 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 459 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-classroom-20-b4-humanscale-nld-eng | 0.5841 | 0.7871 | nan | nan | nan | 0.7403 | nan | nan | nan | nan | nan | nan | 0.8983 | nan | nan | nan | nan | nan |
|
| 460 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-0-100-b3-humanscale-nld-eng | 0.5739 | 0.7871 | nan | nan | nan | 0.7455 | nan | nan | nan | nan | nan | nan | 0.8975 | nan | nan | nan | nan | nan |
|
| 461 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-33-67-b3-humanscale-deu-eng | 0.5743 | nan | nan | 0.8995 | nan | 0.713 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 462 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-33-67-b3-humanscale-nld-eng | 0.5825 | 0.7837 | nan | nan | nan | 0.7377 | nan | nan | nan | nan | nan | nan | 0.9026 | nan | nan | nan | nan | nan |
|
| 463 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-sequential-33-67-b3-humanscale-zho-eng | 0.5472 | nan | nan | nan | nan | 0.7039 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | 0.701 |
|
| 464 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-simultaneous-b2-humanscale-deu-eng | 0.5605 | nan | 43.7057 | 0.8951 | nan | 0.7078 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 465 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-simultaneous-b2-humanscale-nld-eng | 0.5734 | 0.7914 | 44.776 | nan | nan | 0.7403 | nan | nan | nan | nan | nan | nan | 0.888 | nan | nan | nan | nan | nan |
|
| 466 |
+
| Beetle-HumanScale/beetle-bilingual-l2-50-simultaneous-b2-humanscale-zho-eng | 0.5423 | nan | 42.5336 | nan | nan | 0.6883 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | 0.6975 |
|
| 467 |
+
| Beetle-HumanScale/beetle-bilingual-l2-80-late-b5-humanscale-deu-eng | 0.5579 | nan | nan | 0.8956 | nan | 0.7117 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan |
|
| 468 |
+
| Beetle-HumanScale/beetle-bilingual-l2-80-late-b5-humanscale-nld-eng | 0.5664 | 0.7942 | nan | nan | nan | 0.7247 | nan | nan | nan | nan | nan | nan | 0.8846 | nan | nan | nan | nan | nan |
|
| 469 |
+
| Beetle-HumanScale/beetle-bilingual-l2-80-late-b5-humanscale-zho-eng | 0.5489 | nan | nan | nan | nan | 0.6961 | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | nan | 0.7144 |
|
| 470 |
+
| Beetle-HumanScale/beetle-monolingual-humanscale-deu | 0.5115 | 0.4794 | nan | 0.8999 | nan | 0.6468 | nan | nan | nan | nan | nan | nan | 0.4925 | nan | nan | nan | nan | 0.2305 |
|
| 471 |
+
| Beetle-HumanScale/beetle-monolingual-humanscale-eng | 0.6472 | 0.492 | 45.7675 | 0.6023 | nan | 0.7896 | nan | nan | nan | nan | nan | nan | 0.5302 | nan | nan | nan | nan | 0.2327 |
|
| 472 |
+
| Beetle-HumanScale/beetle-monolingual-humanscale-nld | 0.5252 | 0.7897 | nan | 0.6745 | nan | 0.6338 | nan | nan | nan | nan | nan | nan | 0.8885 | nan | nan | nan | nan | 0.1881 |
|
| 473 |
+
| Beetle-HumanScale/beetle-monolingual-humanscale-zho | 0.515 | 0.4788 | nan | 0.5949 | nan | 0.6182 | nan | nan | nan | nan | nan | nan | 0.5109 | nan | nan | nan | nan | 0.7465 |
|
| 474 |
+
| Beetle-nld-eng-Variants/beetle-bilingual-l2-50-sequential-33-67-b3-fineweb-2b-nld-eng-arch-moe | 0.5592 | 0.7638 | nan | nan | nan | 0.7117 | nan | nan | nan | nan | nan | nan | 0.888 | nan | nan | nan | nan | nan |
|
beetle_merge_eval_deltas.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
beetle_merge_eval_long.csv
CHANGED
|
The diff for this file is too large to render.
See raw diff
|
|
|
coverage_by_family.csv
CHANGED
|
@@ -2,5 +2,5 @@ provenance_family,merges,rungs_published,rungs_scored,complete,partial,not_eval
|
|
| 2 |
arch,1,6,6,1,0,0
|
| 3 |
humanscale-pair,3,72,72,3,0,0
|
| 4 |
zoo-curriculum,49,49,49,49,0,0
|
| 5 |
-
zoo-language,111,692,
|
| 6 |
zoo-mono_vs_joint,50,50,50,50,0,0
|
|
|
|
| 2 |
arch,1,6,6,1,0,0
|
| 3 |
humanscale-pair,3,72,72,3,0,0
|
| 4 |
zoo-curriculum,49,49,49,49,0,0
|
| 5 |
+
zoo-language,111,692,692,111,0,0
|
| 6 |
zoo-mono_vs_joint,50,50,50,50,0,0
|
coverage_by_merge.csv
CHANGED
|
@@ -52,115 +52,115 @@ Mergeability/zoo-curriculum-beetle-bilingual-l2-50-simultaneous-b2-humanscale-nl
|
|
| 52 |
Mergeability/zoo-curriculum-beetle-bilingual-l2-50-simultaneous-b2-humanscale-nld-eng__beetle-biling-9499847e,zoo-curriculum,1,1,4,complete
|
| 53 |
Mergeability/zoo-curriculum-beetle-bilingual-l2-50-simultaneous-b2-humanscale-zho-eng__beetle-biling-8d8163db,zoo-curriculum,1,1,3,complete
|
| 54 |
Mergeability/zoo-curriculum-beetle-bilingual-l2-50-simultaneous-b2-humanscale-zho-eng__beetle-biling-d035ff88,zoo-curriculum,1,1,3,complete
|
| 55 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-ell,zoo-language,6,
|
| 56 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-est,zoo-language,6,
|
| 57 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-fin,zoo-language,6,
|
| 58 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-hin,zoo-language,6,
|
| 59 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-isl,zoo-language,6,
|
| 60 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-ita,zoo-language,6,
|
| 61 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-jpn,zoo-language,7,
|
| 62 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 63 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 64 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-pol,zoo-language,7,
|
| 65 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 66 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 67 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 68 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 69 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-est,zoo-language,6,
|
| 70 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-fin,zoo-language,6,
|
| 71 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-hin,zoo-language,6,
|
| 72 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-isl,zoo-language,6,
|
| 73 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-ita,zoo-language,6,
|
| 74 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,
|
| 75 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 76 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 77 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 78 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 79 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 80 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 81 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 82 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-fin,zoo-language,6,
|
| 83 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-hin,zoo-language,6,
|
| 84 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-isl,zoo-language,6,
|
| 85 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-ita,zoo-language,6,
|
| 86 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,
|
| 87 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 88 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 89 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 90 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 91 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 92 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 93 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 94 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-hin,zoo-language,6,
|
| 95 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-isl,zoo-language,6,
|
| 96 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-ita,zoo-language,6,
|
| 97 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,
|
| 98 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 99 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 100 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 101 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 102 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 103 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 104 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 105 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-isl,zoo-language,6,
|
| 106 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-ita,zoo-language,6,
|
| 107 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,
|
| 108 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 109 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 110 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 111 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 112 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 113 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 114 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 115 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-ita,zoo-language,6,
|
| 116 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,
|
| 117 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 118 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 119 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 120 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 121 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 122 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 123 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 124 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,
|
| 125 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 126 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 127 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 128 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 129 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 130 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 131 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 132 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-kor,zoo-language,6,
|
| 133 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 134 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 135 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 136 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 137 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 138 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 139 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-nld,zoo-language,6,
|
| 140 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 141 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 142 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 143 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 144 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 145 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-pol,zoo-language,6,
|
| 146 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 147 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 148 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 149 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 150 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-rus,zoo-language,6,
|
| 151 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 152 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 153 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 154 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-rus__beetle-monolingual-fineweb-100m-spa,zoo-language,6,
|
| 155 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-rus__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 156 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-rus__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 157 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-spa__beetle-monolingual-fineweb-100m-tur,zoo-language,6,
|
| 158 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-spa__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 159 |
-
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-tur__beetle-monolingual-fineweb-100m-zho,zoo-language,6,
|
| 160 |
Mergeability/zoo-language-beetle-monolingual-humanscale-deu__beetle-monolingual-humanscale-eng,zoo-language,10,10,30,complete
|
| 161 |
Mergeability/zoo-language-beetle-monolingual-humanscale-deu__beetle-monolingual-humanscale-nld,zoo-language,10,10,50,complete
|
| 162 |
-
Mergeability/zoo-language-beetle-monolingual-humanscale-deu__beetle-monolingual-humanscale-zho,zoo-language,10,
|
| 163 |
-
Mergeability/zoo-language-beetle-monolingual-humanscale-eng__beetle-monolingual-humanscale-nld,zoo-language,10,
|
| 164 |
Mergeability/zoo-language-beetle-monolingual-humanscale-eng__beetle-monolingual-humanscale-zho,zoo-language,10,10,30,complete
|
| 165 |
Mergeability/zoo-language-beetle-monolingual-humanscale-nld__beetle-monolingual-humanscale-zho,zoo-language,10,10,50,complete
|
| 166 |
Mergeability/zoo-mono_vs_joint-beetle-monolingual-fineweb-100m-deu__beetle-bilingual-balanced-b1-fin-57288df9,zoo-mono_vs_joint,1,1,3,complete
|
|
|
|
| 52 |
Mergeability/zoo-curriculum-beetle-bilingual-l2-50-simultaneous-b2-humanscale-nld-eng__beetle-biling-9499847e,zoo-curriculum,1,1,4,complete
|
| 53 |
Mergeability/zoo-curriculum-beetle-bilingual-l2-50-simultaneous-b2-humanscale-zho-eng__beetle-biling-8d8163db,zoo-curriculum,1,1,3,complete
|
| 54 |
Mergeability/zoo-curriculum-beetle-bilingual-l2-50-simultaneous-b2-humanscale-zho-eng__beetle-biling-d035ff88,zoo-curriculum,1,1,3,complete
|
| 55 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-ell,zoo-language,6,6,24,complete
|
| 56 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-est,zoo-language,6,6,12,complete
|
| 57 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-fin,zoo-language,6,6,24,complete
|
| 58 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-hin,zoo-language,6,6,24,complete
|
| 59 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-isl,zoo-language,6,6,24,complete
|
| 60 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-ita,zoo-language,6,6,24,complete
|
| 61 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-jpn,zoo-language,7,7,21,complete
|
| 62 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,18,complete
|
| 63 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,30,complete
|
| 64 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-pol,zoo-language,7,7,28,complete
|
| 65 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,24,complete
|
| 66 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 67 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 68 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-deu__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 69 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-est,zoo-language,6,6,12,complete
|
| 70 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-fin,zoo-language,6,6,24,complete
|
| 71 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-hin,zoo-language,6,6,24,complete
|
| 72 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-isl,zoo-language,6,6,24,complete
|
| 73 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-ita,zoo-language,6,6,24,complete
|
| 74 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,6,18,complete
|
| 75 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,18,complete
|
| 76 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,30,complete
|
| 77 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,24,complete
|
| 78 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,24,complete
|
| 79 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 80 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 81 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ell__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 82 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-fin,zoo-language,6,6,12,complete
|
| 83 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-hin,zoo-language,6,6,12,complete
|
| 84 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-isl,zoo-language,6,6,12,complete
|
| 85 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-ita,zoo-language,6,6,12,complete
|
| 86 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,6,12,complete
|
| 87 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,12,complete
|
| 88 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,12,complete
|
| 89 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,12,complete
|
| 90 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,12,complete
|
| 91 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,12,complete
|
| 92 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,12,complete
|
| 93 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-est__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,12,complete
|
| 94 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-hin,zoo-language,6,6,24,complete
|
| 95 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-isl,zoo-language,6,6,24,complete
|
| 96 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-ita,zoo-language,6,6,24,complete
|
| 97 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,6,18,complete
|
| 98 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,18,complete
|
| 99 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,30,complete
|
| 100 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,24,complete
|
| 101 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,24,complete
|
| 102 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 103 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 104 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-fin__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 105 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-isl,zoo-language,6,6,24,complete
|
| 106 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-ita,zoo-language,6,6,24,complete
|
| 107 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,6,18,complete
|
| 108 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,18,complete
|
| 109 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,30,complete
|
| 110 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,24,complete
|
| 111 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,24,complete
|
| 112 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 113 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 114 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-hin__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 115 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-ita,zoo-language,6,6,24,complete
|
| 116 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,6,18,complete
|
| 117 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,18,complete
|
| 118 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,30,complete
|
| 119 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,24,complete
|
| 120 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,24,complete
|
| 121 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 122 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 123 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-isl__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 124 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-jpn,zoo-language,6,6,18,complete
|
| 125 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,18,complete
|
| 126 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,30,complete
|
| 127 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,24,complete
|
| 128 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,24,complete
|
| 129 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 130 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 131 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-ita__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 132 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-kor,zoo-language,6,6,12,complete
|
| 133 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,24,complete
|
| 134 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,18,complete
|
| 135 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,18,complete
|
| 136 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,18,complete
|
| 137 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,18,complete
|
| 138 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-jpn__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,18,complete
|
| 139 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-nld,zoo-language,6,6,24,complete
|
| 140 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,18,complete
|
| 141 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,18,complete
|
| 142 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,18,complete
|
| 143 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,18,complete
|
| 144 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-kor__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,18,complete
|
| 145 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-pol,zoo-language,6,6,30,complete
|
| 146 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,30,complete
|
| 147 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,30,complete
|
| 148 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,30,complete
|
| 149 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-nld__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,30,complete
|
| 150 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-rus,zoo-language,6,6,24,complete
|
| 151 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 152 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 153 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-pol__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 154 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-rus__beetle-monolingual-fineweb-100m-spa,zoo-language,6,6,24,complete
|
| 155 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-rus__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 156 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-rus__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 157 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-spa__beetle-monolingual-fineweb-100m-tur,zoo-language,6,6,24,complete
|
| 158 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-spa__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 159 |
+
Mergeability/zoo-language-beetle-monolingual-fineweb-100m-tur__beetle-monolingual-fineweb-100m-zho,zoo-language,6,6,24,complete
|
| 160 |
Mergeability/zoo-language-beetle-monolingual-humanscale-deu__beetle-monolingual-humanscale-eng,zoo-language,10,10,30,complete
|
| 161 |
Mergeability/zoo-language-beetle-monolingual-humanscale-deu__beetle-monolingual-humanscale-nld,zoo-language,10,10,50,complete
|
| 162 |
+
Mergeability/zoo-language-beetle-monolingual-humanscale-deu__beetle-monolingual-humanscale-zho,zoo-language,10,10,40,complete
|
| 163 |
+
Mergeability/zoo-language-beetle-monolingual-humanscale-eng__beetle-monolingual-humanscale-nld,zoo-language,10,10,40,complete
|
| 164 |
Mergeability/zoo-language-beetle-monolingual-humanscale-eng__beetle-monolingual-humanscale-zho,zoo-language,10,10,30,complete
|
| 165 |
Mergeability/zoo-language-beetle-monolingual-humanscale-nld__beetle-monolingual-humanscale-zho,zoo-language,10,10,50,complete
|
| 166 |
Mergeability/zoo-mono_vs_joint-beetle-monolingual-fineweb-100m-deu__beetle-bilingual-balanced-b1-fin-57288df9,zoo-mono_vs_joint,1,1,3,complete
|
fig_damage_vs_parent_relatedness.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
fig_headline_naive_vs_aligned.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
fig_merge_vs_joint_ceiling.png
CHANGED
|
Git LFS Details
|
|
Git LFS Details
|
summary_by_family_arm.csv
CHANGED
|
@@ -5,8 +5,8 @@ humanscale-pair,aligned,80.0,3.0,0.5614,0.0875,0.2,0.0,0.0,80.0,-0.2021,0.0,80.0
|
|
| 5 |
humanscale-pair,naive,80.0,3.0,0.5614,0.0875,0.2,0.0,0.0,80.0,-0.2021,0.0,80.0,-0.1471,1.0,80.0
|
| 6 |
humanscale-pair,transport,80.0,3.0,0.5779,0.1,0.2125,0.0165,42.0,80.0,-0.1856,0.0,80.0,-0.1305,2.0,80.0
|
| 7 |
zoo-curriculum,aligned,158.0,49.0,0.739,0.0,0.0,,0.0,0.0,-0.0184,16.0,158.0,,0.0,0.0
|
| 8 |
-
zoo-language,aligned,
|
| 9 |
-
zoo-language,naive,
|
| 10 |
-
zoo-language,transport,
|
| 11 |
zoo-mono_vs_joint,aligned,70.0,22.0,0.6645,0.0143,0.0429,,0.0,0.0,-0.0663,1.0,70.0,-0.0653,2.0,70.0
|
| 12 |
-
zoo-mono_vs_joint,transport,96.0,28.0,0.
|
|
|
|
| 5 |
humanscale-pair,naive,80.0,3.0,0.5614,0.0875,0.2,0.0,0.0,80.0,-0.2021,0.0,80.0,-0.1471,1.0,80.0
|
| 6 |
humanscale-pair,transport,80.0,3.0,0.5779,0.1,0.2125,0.0165,42.0,80.0,-0.1856,0.0,80.0,-0.1305,2.0,80.0
|
| 7 |
zoo-curriculum,aligned,158.0,49.0,0.739,0.0,0.0,,0.0,0.0,-0.0184,16.0,158.0,,0.0,0.0
|
| 8 |
+
zoo-language,aligned,451.0,111.0,0.5883,0.0665,0.1752,0.0,6.0,451.0,-0.1241,53.0,451.0,,0.0,0.0
|
| 9 |
+
zoo-language,naive,458.0,111.0,0.5895,0.0655,0.1725,0.0,0.0,458.0,-0.1229,55.0,458.0,,0.0,0.0
|
| 10 |
+
zoo-language,transport,1612.0,111.0,0.6186,0.0509,0.1569,0.0056,260.0,458.0,-0.0904,224.0,1612.0,,0.0,0.0
|
| 11 |
zoo-mono_vs_joint,aligned,70.0,22.0,0.6645,0.0143,0.0429,,0.0,0.0,-0.0663,1.0,70.0,-0.0653,2.0,70.0
|
| 12 |
+
zoo-mono_vs_joint,transport,96.0,28.0,0.6537,0.0625,0.1458,,0.0,0.0,-0.0968,10.0,96.0,-0.0755,34.0,96.0
|
summary_by_family_bench_arm.csv
CHANGED
|
@@ -26,63 +26,75 @@ humanscale-pair,zhoblimp,aligned,8.0,1.0,0.5813,0.0,0.0,0.0,0.0,8.0,-0.1652,0.0,
|
|
| 26 |
humanscale-pair,zhoblimp,naive,8.0,1.0,0.5813,0.0,0.0,0.0,0.0,8.0,-0.1652,0.0,8.0,-0.1162,1.0,8.0
|
| 27 |
humanscale-pair,zhoblimp,transport,8.0,1.0,0.5872,0.0,0.0,0.0059,3.0,8.0,-0.1593,0.0,8.0,-0.1103,1.0,8.0
|
| 28 |
zoo-curriculum,blimp_eng,aligned,49.0,49.0,0.5925,0.0,0.0,,0.0,0.0,-0.0137,5.0,49.0,,0.0,0.0
|
| 29 |
-
zoo-curriculum,blimp_nl,aligned,14.0,14.0,0.
|
| 30 |
zoo-curriculum,multiblimp:deu,aligned,11.0,11.0,0.8856,0.0,0.0,,0.0,0.0,-0.0179,2.0,11.0,,0.0,0.0
|
| 31 |
zoo-curriculum,multiblimp:eng,aligned,49.0,49.0,0.7711,0.0,0.0,,0.0,0.0,-0.0194,4.0,49.0,,0.0,0.0
|
| 32 |
zoo-curriculum,multiblimp:eus,aligned,3.0,3.0,0.978,0.0,0.0,,0.0,0.0,0.0037,2.0,3.0,,0.0,0.0
|
| 33 |
zoo-curriculum,multiblimp:hin,aligned,3.0,3.0,0.8659,0.0,0.0,,0.0,0.0,-0.0514,0.0,3.0,,0.0,0.0
|
| 34 |
zoo-curriculum,multiblimp:ita,aligned,3.0,3.0,0.8142,0.0,0.0,,0.0,0.0,-0.0343,0.0,3.0,,0.0,0.0
|
| 35 |
-
zoo-curriculum,multiblimp:nld,aligned,14.0,14.0,0.8828,0.0,0.0,,0.0,0.0,-0.
|
| 36 |
zoo-curriculum,multiblimp:pol,aligned,3.0,3.0,0.7879,0.0,0.0,,0.0,0.0,-0.0329,0.0,3.0,,0.0,0.0
|
| 37 |
-
zoo-curriculum,multiblimp:rus,aligned,3.0,3.0,0.
|
| 38 |
zoo-curriculum,multiblimp:tur,aligned,3.0,3.0,0.7773,0.0,0.0,,0.0,0.0,-0.0318,0.0,3.0,,0.0,0.0
|
| 39 |
-
zoo-curriculum,zhoblimp,aligned,3.0,3.0,0.7221,0.0,0.0,,0.0,0.0,0.
|
| 40 |
-
zoo-language,blimp_eng,aligned,
|
| 41 |
-
zoo-language,blimp_eng,naive,
|
| 42 |
-
zoo-language,blimp_eng,transport,
|
| 43 |
-
zoo-language,blimp_nl,aligned,
|
| 44 |
-
zoo-language,blimp_nl,naive,
|
| 45 |
-
zoo-language,blimp_nl,transport,
|
| 46 |
-
zoo-language,multiblimp:deu,aligned,
|
| 47 |
-
zoo-language,multiblimp:deu,naive,
|
| 48 |
-
zoo-language,multiblimp:deu,transport,
|
| 49 |
-
zoo-language,multiblimp:ell,aligned,
|
| 50 |
-
zoo-language,multiblimp:ell,
|
| 51 |
-
zoo-language,multiblimp:
|
| 52 |
-
zoo-language,multiblimp:eng,
|
| 53 |
-
zoo-language,multiblimp:eng,
|
| 54 |
-
zoo-language,multiblimp:
|
| 55 |
-
zoo-language,multiblimp:fin,
|
| 56 |
-
zoo-language,multiblimp:
|
| 57 |
-
zoo-language,multiblimp:
|
| 58 |
-
zoo-language,multiblimp:
|
| 59 |
-
zoo-language,multiblimp:
|
| 60 |
-
zoo-language,multiblimp:
|
| 61 |
-
zoo-language,multiblimp:
|
| 62 |
-
zoo-language,multiblimp:
|
| 63 |
-
zoo-language,multiblimp:
|
| 64 |
-
zoo-language,multiblimp:
|
| 65 |
-
zoo-language,multiblimp:
|
| 66 |
-
zoo-language,multiblimp:
|
| 67 |
-
zoo-language,multiblimp:
|
| 68 |
-
zoo-language,multiblimp:
|
| 69 |
-
zoo-language,multiblimp:
|
| 70 |
-
zoo-language,
|
| 71 |
-
zoo-language,
|
| 72 |
-
zoo-language,
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 73 |
zoo-mono_vs_joint,blimp_eng,aligned,22.0,22.0,0.5314,0.0455,0.1364,,0.0,0.0,-0.0418,1.0,22.0,-0.0407,2.0,22.0
|
| 74 |
-
zoo-mono_vs_joint,blimp_eng,transport,28.0,28.0,0.
|
| 75 |
-
zoo-mono_vs_joint,blimp_nl,aligned,4.0,4.0,0.7287,0.0,0.0,,0.0,0.0,-0.
|
| 76 |
-
zoo-mono_vs_joint,blimp_nl,transport,12.0,12.0,0.6438,0.5,0.5,,0.0,0.0,-0.1442,
|
| 77 |
zoo-mono_vs_joint,multiblimp:deu,aligned,3.0,3.0,0.8639,0.0,0.0,,0.0,0.0,-0.0461,0.0,3.0,-0.0441,0.0,3.0
|
| 78 |
-
zoo-mono_vs_joint,multiblimp:deu,transport,10.0,10.0,0.
|
| 79 |
-
zoo-mono_vs_joint,multiblimp:eng,aligned,22.0,22.0,0.
|
| 80 |
zoo-mono_vs_joint,multiblimp:eng,transport,28.0,28.0,0.7078,0.0,0.0,,0.0,0.0,-0.054,3.0,28.0,-0.0239,12.0,28.0
|
| 81 |
zoo-mono_vs_joint,multiblimp:hin,aligned,3.0,3.0,0.8123,0.0,0.0,,0.0,0.0,-0.091,0.0,3.0,-0.091,0.0,3.0
|
| 82 |
zoo-mono_vs_joint,multiblimp:ita,aligned,3.0,3.0,0.7892,0.0,0.0,,0.0,0.0,-0.0488,0.0,3.0,-0.0488,0.0,3.0
|
| 83 |
-
zoo-mono_vs_joint,multiblimp:nld,aligned,4.0,4.0,0.
|
| 84 |
zoo-mono_vs_joint,multiblimp:nld,transport,12.0,12.0,0.7094,0.0,0.0,,0.0,0.0,-0.1831,0.0,12.0,-0.1816,2.0,12.0
|
| 85 |
zoo-mono_vs_joint,multiblimp:pol,aligned,3.0,3.0,0.768,0.0,0.0,,0.0,0.0,-0.0407,0.0,3.0,-0.0372,0.0,3.0
|
| 86 |
zoo-mono_vs_joint,multiblimp:rus,aligned,3.0,3.0,0.757,0.0,0.0,,0.0,0.0,-0.0522,0.0,3.0,-0.0456,0.0,3.0
|
| 87 |
zoo-mono_vs_joint,multiblimp:tur,aligned,3.0,3.0,0.7512,0.0,0.0,,0.0,0.0,-0.0431,0.0,3.0,-0.0425,0.0,3.0
|
| 88 |
-
zoo-mono_vs_joint,zhoblimp,transport,6.0,6.0,0.
|
|
|
|
| 26 |
humanscale-pair,zhoblimp,naive,8.0,1.0,0.5813,0.0,0.0,0.0,0.0,8.0,-0.1652,0.0,8.0,-0.1162,1.0,8.0
|
| 27 |
humanscale-pair,zhoblimp,transport,8.0,1.0,0.5872,0.0,0.0,0.0059,3.0,8.0,-0.1593,0.0,8.0,-0.1103,1.0,8.0
|
| 28 |
zoo-curriculum,blimp_eng,aligned,49.0,49.0,0.5925,0.0,0.0,,0.0,0.0,-0.0137,5.0,49.0,,0.0,0.0
|
| 29 |
+
zoo-curriculum,blimp_nl,aligned,14.0,14.0,0.765,0.0,0.0,,0.0,0.0,-0.0158,0.0,14.0,,0.0,0.0
|
| 30 |
zoo-curriculum,multiblimp:deu,aligned,11.0,11.0,0.8856,0.0,0.0,,0.0,0.0,-0.0179,2.0,11.0,,0.0,0.0
|
| 31 |
zoo-curriculum,multiblimp:eng,aligned,49.0,49.0,0.7711,0.0,0.0,,0.0,0.0,-0.0194,4.0,49.0,,0.0,0.0
|
| 32 |
zoo-curriculum,multiblimp:eus,aligned,3.0,3.0,0.978,0.0,0.0,,0.0,0.0,0.0037,2.0,3.0,,0.0,0.0
|
| 33 |
zoo-curriculum,multiblimp:hin,aligned,3.0,3.0,0.8659,0.0,0.0,,0.0,0.0,-0.0514,0.0,3.0,,0.0,0.0
|
| 34 |
zoo-curriculum,multiblimp:ita,aligned,3.0,3.0,0.8142,0.0,0.0,,0.0,0.0,-0.0343,0.0,3.0,,0.0,0.0
|
| 35 |
+
zoo-curriculum,multiblimp:nld,aligned,14.0,14.0,0.8828,0.0,0.0,,0.0,0.0,-0.0211,0.0,14.0,,0.0,0.0
|
| 36 |
zoo-curriculum,multiblimp:pol,aligned,3.0,3.0,0.7879,0.0,0.0,,0.0,0.0,-0.0329,0.0,3.0,,0.0,0.0
|
| 37 |
+
zoo-curriculum,multiblimp:rus,aligned,3.0,3.0,0.765,0.0,0.0,,0.0,0.0,-0.0531,0.0,3.0,,0.0,0.0
|
| 38 |
zoo-curriculum,multiblimp:tur,aligned,3.0,3.0,0.7773,0.0,0.0,,0.0,0.0,-0.0318,0.0,3.0,,0.0,0.0
|
| 39 |
+
zoo-curriculum,zhoblimp,aligned,3.0,3.0,0.7221,0.0,0.0,,0.0,0.0,0.0122,3.0,3.0,,0.0,0.0
|
| 40 |
+
zoo-language,blimp_eng,aligned,123.0,111.0,0.5234,0.0813,0.439,-0.0,2.0,123.0,-0.0209,26.0,123.0,,0.0,0.0
|
| 41 |
+
zoo-language,blimp_eng,naive,125.0,111.0,0.5235,0.08,0.432,0.0,0.0,125.0,-0.0206,27.0,125.0,,0.0,0.0
|
| 42 |
+
zoo-language,blimp_eng,transport,444.0,111.0,0.5257,0.0405,0.3559,0.0008,67.0,125.0,-0.0143,93.0,444.0,,0.0,0.0
|
| 43 |
+
zoo-language,blimp_nl,aligned,22.0,16.0,0.5123,0.2273,0.4091,-0.0,0.0,22.0,-0.2448,0.0,22.0,,0.0,0.0
|
| 44 |
+
zoo-language,blimp_nl,naive,22.0,16.0,0.5123,0.2273,0.4091,0.0,0.0,22.0,-0.2448,0.0,22.0,,0.0,0.0
|
| 45 |
+
zoo-language,blimp_nl,transport,64.0,16.0,0.5413,0.1562,0.375,0.0168,16.0,22.0,-0.2035,0.0,64.0,,0.0,0.0
|
| 46 |
+
zoo-language,multiblimp:deu,aligned,22.0,16.0,0.7355,0.0,0.0,0.0,1.0,22.0,-0.16,0.0,22.0,,0.0,0.0
|
| 47 |
+
zoo-language,multiblimp:deu,naive,24.0,16.0,0.7433,0.0,0.0,0.0,0.0,24.0,-0.152,0.0,24.0,,0.0,0.0
|
| 48 |
+
zoo-language,multiblimp:deu,transport,64.0,16.0,0.7797,0.0,0.0,-0.0091,6.0,24.0,-0.1142,0.0,64.0,,0.0,0.0
|
| 49 |
+
zoo-language,multiblimp:ell,aligned,13.0,13.0,0.6779,0.0,0.0,0.0,0.0,13.0,-0.2254,0.0,13.0,,0.0,0.0
|
| 50 |
+
zoo-language,multiblimp:ell,naive,13.0,13.0,0.6779,0.0,0.0,0.0,0.0,13.0,-0.2254,0.0,13.0,,0.0,0.0
|
| 51 |
+
zoo-language,multiblimp:ell,transport,52.0,13.0,0.8071,0.0,0.0,0.0361,12.0,13.0,-0.0962,0.0,52.0,,0.0,0.0
|
| 52 |
+
zoo-language,multiblimp:eng,aligned,123.0,111.0,0.6598,0.0,0.0,0.0,1.0,123.0,-0.029,27.0,123.0,,0.0,0.0
|
| 53 |
+
zoo-language,multiblimp:eng,naive,125.0,111.0,0.66,0.0,0.0,0.0,0.0,125.0,-0.0285,28.0,125.0,,0.0,0.0
|
| 54 |
+
zoo-language,multiblimp:eng,transport,444.0,111.0,0.6688,0.0023,0.0,0.004,65.0,125.0,-0.017,131.0,444.0,,0.0,0.0
|
| 55 |
+
zoo-language,multiblimp:fin,aligned,13.0,13.0,0.6197,0.0,0.0,0.0,0.0,13.0,-0.2192,0.0,13.0,,0.0,0.0
|
| 56 |
+
zoo-language,multiblimp:fin,naive,13.0,13.0,0.6197,0.0,0.0,0.0,0.0,13.0,-0.2192,0.0,13.0,,0.0,0.0
|
| 57 |
+
zoo-language,multiblimp:fin,transport,52.0,13.0,0.7026,0.0,0.0,0.0047,7.0,13.0,-0.1364,0.0,52.0,,0.0,0.0
|
| 58 |
+
zoo-language,multiblimp:hin,aligned,13.0,13.0,0.5493,0.3077,0.1538,0.0,0.0,13.0,-0.325,0.0,13.0,,0.0,0.0
|
| 59 |
+
zoo-language,multiblimp:hin,naive,13.0,13.0,0.5493,0.3077,0.1538,0.0,0.0,13.0,-0.325,0.0,13.0,,0.0,0.0
|
| 60 |
+
zoo-language,multiblimp:hin,transport,52.0,13.0,0.6536,0.1538,0.1346,-0.0007,7.0,13.0,-0.2207,0.0,52.0,,0.0,0.0
|
| 61 |
+
zoo-language,multiblimp:isl,aligned,13.0,13.0,0.664,0.0,0.0,0.0,0.0,13.0,-0.2564,0.0,13.0,,0.0,0.0
|
| 62 |
+
zoo-language,multiblimp:isl,naive,13.0,13.0,0.664,0.0,0.0,0.0,0.0,13.0,-0.2564,0.0,13.0,,0.0,0.0
|
| 63 |
+
zoo-language,multiblimp:isl,transport,52.0,13.0,0.7508,0.0,0.0,0.0018,8.0,13.0,-0.1696,0.0,52.0,,0.0,0.0
|
| 64 |
+
zoo-language,multiblimp:ita,aligned,13.0,13.0,0.5547,0.0769,0.0769,0.0,0.0,13.0,-0.2542,0.0,13.0,,0.0,0.0
|
| 65 |
+
zoo-language,multiblimp:ita,naive,13.0,13.0,0.5547,0.0769,0.0769,0.0,0.0,13.0,-0.2542,0.0,13.0,,0.0,0.0
|
| 66 |
+
zoo-language,multiblimp:ita,transport,52.0,13.0,0.6208,0.2692,0.2885,0.0209,12.0,13.0,-0.1881,0.0,52.0,,0.0,0.0
|
| 67 |
+
zoo-language,multiblimp:nld,aligned,22.0,16.0,0.5617,0.2273,0.1818,0.0,0.0,22.0,-0.3225,0.0,22.0,,0.0,0.0
|
| 68 |
+
zoo-language,multiblimp:nld,naive,22.0,16.0,0.5617,0.2273,0.1818,0.0,0.0,22.0,-0.3225,0.0,22.0,,0.0,0.0
|
| 69 |
+
zoo-language,multiblimp:nld,transport,64.0,16.0,0.595,0.2188,0.2188,0.015,12.0,22.0,-0.2875,0.0,64.0,,0.0,0.0
|
| 70 |
+
zoo-language,multiblimp:pol,aligned,13.0,13.0,0.6023,0.0,0.0,0.0,0.0,13.0,-0.1859,0.0,13.0,,0.0,0.0
|
| 71 |
+
zoo-language,multiblimp:pol,naive,14.0,13.0,0.6019,0.0,0.0,0.0,0.0,14.0,-0.1863,0.0,14.0,,0.0,0.0
|
| 72 |
+
zoo-language,multiblimp:pol,transport,52.0,13.0,0.634,0.0,0.0,0.0144,12.0,14.0,-0.1542,0.0,52.0,,0.0,0.0
|
| 73 |
+
zoo-language,multiblimp:rus,aligned,13.0,13.0,0.538,0.0769,0.1538,0.0,0.0,13.0,-0.2561,0.0,13.0,,0.0,0.0
|
| 74 |
+
zoo-language,multiblimp:rus,naive,13.0,13.0,0.538,0.0769,0.1538,0.0,0.0,13.0,-0.2561,0.0,13.0,,0.0,0.0
|
| 75 |
+
zoo-language,multiblimp:rus,transport,52.0,13.0,0.5683,0.1154,0.1346,0.0193,12.0,13.0,-0.2258,0.0,52.0,,0.0,0.0
|
| 76 |
+
zoo-language,multiblimp:spa,aligned,13.0,13.0,0.6657,0.0,0.0,0.0,0.0,13.0,-0.232,0.0,13.0,,0.0,0.0
|
| 77 |
+
zoo-language,multiblimp:spa,naive,13.0,13.0,0.6657,0.0,0.0,0.0,0.0,13.0,-0.232,0.0,13.0,,0.0,0.0
|
| 78 |
+
zoo-language,multiblimp:spa,transport,52.0,13.0,0.6898,0.0,0.0,0.0165,9.0,13.0,-0.2079,0.0,52.0,,0.0,0.0
|
| 79 |
+
zoo-language,multiblimp:tur,aligned,13.0,13.0,0.6438,0.0,0.0,0.0,1.0,13.0,-0.1225,0.0,13.0,,0.0,0.0
|
| 80 |
+
zoo-language,multiblimp:tur,naive,13.0,13.0,0.6438,0.0,0.0,0.0,0.0,13.0,-0.1226,0.0,13.0,,0.0,0.0
|
| 81 |
+
zoo-language,multiblimp:tur,transport,52.0,13.0,0.6538,0.0,0.0,0.0047,7.0,13.0,-0.1125,0.0,52.0,,0.0,0.0
|
| 82 |
+
zoo-language,zhoblimp,aligned,22.0,16.0,0.3765,0.1818,0.3182,-0.0001,1.0,22.0,-0.3105,0.0,22.0,,0.0,0.0
|
| 83 |
+
zoo-language,zhoblimp,naive,22.0,16.0,0.3766,0.1818,0.3182,0.0,0.0,22.0,-0.3104,0.0,22.0,,0.0,0.0
|
| 84 |
+
zoo-language,zhoblimp,transport,64.0,16.0,0.4373,0.1719,0.4375,-0.0034,8.0,22.0,-0.2275,0.0,64.0,,0.0,0.0
|
| 85 |
zoo-mono_vs_joint,blimp_eng,aligned,22.0,22.0,0.5314,0.0455,0.1364,,0.0,0.0,-0.0418,1.0,22.0,-0.0407,2.0,22.0
|
| 86 |
+
zoo-mono_vs_joint,blimp_eng,transport,28.0,28.0,0.5827,0.0,0.2857,,0.0,0.0,-0.0329,5.0,28.0,0.0023,12.0,28.0
|
| 87 |
+
zoo-mono_vs_joint,blimp_nl,aligned,4.0,4.0,0.7287,0.0,0.0,,0.0,0.0,-0.0284,0.0,4.0,-0.0284,0.0,4.0
|
| 88 |
+
zoo-mono_vs_joint,blimp_nl,transport,12.0,12.0,0.6438,0.5,0.5,,0.0,0.0,-0.1442,2.0,12.0,-0.1415,4.0,12.0
|
| 89 |
zoo-mono_vs_joint,multiblimp:deu,aligned,3.0,3.0,0.8639,0.0,0.0,,0.0,0.0,-0.0461,0.0,3.0,-0.0441,0.0,3.0
|
| 90 |
+
zoo-mono_vs_joint,multiblimp:deu,transport,10.0,10.0,0.745,0.0,0.0,,0.0,0.0,-0.1507,0.0,10.0,-0.1464,1.0,10.0
|
| 91 |
+
zoo-mono_vs_joint,multiblimp:eng,aligned,22.0,22.0,0.6503,0.0,0.0,,0.0,0.0,-0.1081,0.0,22.0,-0.1079,0.0,22.0
|
| 92 |
zoo-mono_vs_joint,multiblimp:eng,transport,28.0,28.0,0.7078,0.0,0.0,,0.0,0.0,-0.054,3.0,28.0,-0.0239,12.0,28.0
|
| 93 |
zoo-mono_vs_joint,multiblimp:hin,aligned,3.0,3.0,0.8123,0.0,0.0,,0.0,0.0,-0.091,0.0,3.0,-0.091,0.0,3.0
|
| 94 |
zoo-mono_vs_joint,multiblimp:ita,aligned,3.0,3.0,0.7892,0.0,0.0,,0.0,0.0,-0.0488,0.0,3.0,-0.0488,0.0,3.0
|
| 95 |
+
zoo-mono_vs_joint,multiblimp:nld,aligned,4.0,4.0,0.8446,0.0,0.0,,0.0,0.0,-0.065,0.0,4.0,-0.065,0.0,4.0
|
| 96 |
zoo-mono_vs_joint,multiblimp:nld,transport,12.0,12.0,0.7094,0.0,0.0,,0.0,0.0,-0.1831,0.0,12.0,-0.1816,2.0,12.0
|
| 97 |
zoo-mono_vs_joint,multiblimp:pol,aligned,3.0,3.0,0.768,0.0,0.0,,0.0,0.0,-0.0407,0.0,3.0,-0.0372,0.0,3.0
|
| 98 |
zoo-mono_vs_joint,multiblimp:rus,aligned,3.0,3.0,0.757,0.0,0.0,,0.0,0.0,-0.0522,0.0,3.0,-0.0456,0.0,3.0
|
| 99 |
zoo-mono_vs_joint,multiblimp:tur,aligned,3.0,3.0,0.7512,0.0,0.0,,0.0,0.0,-0.0431,0.0,3.0,-0.0425,0.0,3.0
|
| 100 |
+
zoo-mono_vs_joint,zhoblimp,transport,6.0,6.0,0.4881,0.0,0.0,,0.0,0.0,-0.2373,0.0,6.0,-0.2163,3.0,6.0
|
summary_by_family_benchmark.csv
CHANGED
|
@@ -10,38 +10,40 @@ humanscale-pair,multiblimp:eng,72.0,3.0,0.5609,0.0972,0.0833,0.0093,13.0,72.0,-0
|
|
| 10 |
humanscale-pair,multiblimp:nld,24.0,1.0,0.6042,0.125,0.125,0.0081,4.0,24.0,-0.2842,0.0,24.0,-0.2838,0.0,24.0
|
| 11 |
humanscale-pair,zhoblimp,24.0,1.0,0.5832,0.0,0.0,0.002,3.0,24.0,-0.1633,0.0,24.0,-0.1142,3.0,24.0
|
| 12 |
zoo-curriculum,blimp_eng,49.0,49.0,0.5925,0.0,0.0,,0.0,0.0,-0.0137,5.0,49.0,,0.0,0.0
|
| 13 |
-
zoo-curriculum,blimp_nl,14.0,14.0,0.
|
| 14 |
zoo-curriculum,multiblimp:deu,11.0,11.0,0.8856,0.0,0.0,,0.0,0.0,-0.0179,2.0,11.0,,0.0,0.0
|
| 15 |
zoo-curriculum,multiblimp:eng,49.0,49.0,0.7711,0.0,0.0,,0.0,0.0,-0.0194,4.0,49.0,,0.0,0.0
|
| 16 |
zoo-curriculum,multiblimp:eus,3.0,3.0,0.978,0.0,0.0,,0.0,0.0,0.0037,2.0,3.0,,0.0,0.0
|
| 17 |
zoo-curriculum,multiblimp:hin,3.0,3.0,0.8659,0.0,0.0,,0.0,0.0,-0.0514,0.0,3.0,,0.0,0.0
|
| 18 |
zoo-curriculum,multiblimp:ita,3.0,3.0,0.8142,0.0,0.0,,0.0,0.0,-0.0343,0.0,3.0,,0.0,0.0
|
| 19 |
-
zoo-curriculum,multiblimp:nld,14.0,14.0,0.8828,0.0,0.0,,0.0,0.0,-0.
|
| 20 |
zoo-curriculum,multiblimp:pol,3.0,3.0,0.7879,0.0,0.0,,0.0,0.0,-0.0329,0.0,3.0,,0.0,0.0
|
| 21 |
-
zoo-curriculum,multiblimp:rus,3.0,3.0,0.
|
| 22 |
zoo-curriculum,multiblimp:tur,3.0,3.0,0.7773,0.0,0.0,,0.0,0.0,-0.0318,0.0,3.0,,0.0,0.0
|
| 23 |
-
zoo-curriculum,zhoblimp,3.0,3.0,0.7221,0.0,0.0,,0.0,0.0,0.
|
| 24 |
-
zoo-language,blimp_eng,
|
| 25 |
-
zoo-language,blimp_nl,
|
| 26 |
-
zoo-language,multiblimp:deu,
|
| 27 |
-
zoo-language,multiblimp:ell,
|
| 28 |
-
zoo-language,multiblimp:eng,
|
| 29 |
-
zoo-language,multiblimp:fin,
|
| 30 |
-
zoo-language,multiblimp:hin,
|
| 31 |
-
zoo-language,multiblimp:isl,
|
| 32 |
-
zoo-language,multiblimp:ita,
|
| 33 |
-
zoo-language,multiblimp:nld,
|
| 34 |
-
zoo-language,multiblimp:
|
| 35 |
-
zoo-language,multiblimp:
|
| 36 |
-
zoo-language,
|
|
|
|
|
|
|
| 37 |
zoo-mono_vs_joint,blimp_eng,50.0,50.0,0.5602,0.02,0.22,,0.0,0.0,-0.0368,6.0,50.0,-0.0166,14.0,50.0
|
| 38 |
-
zoo-mono_vs_joint,blimp_nl,16.0,16.0,0.
|
| 39 |
-
zoo-mono_vs_joint,multiblimp:deu,13.0,13.0,0.
|
| 40 |
-
zoo-mono_vs_joint,multiblimp:eng,50.0,50.0,0.6825,0.0,0.0,,0.0,0.0,-0.
|
| 41 |
zoo-mono_vs_joint,multiblimp:hin,3.0,3.0,0.8123,0.0,0.0,,0.0,0.0,-0.091,0.0,3.0,-0.091,0.0,3.0
|
| 42 |
zoo-mono_vs_joint,multiblimp:ita,3.0,3.0,0.7892,0.0,0.0,,0.0,0.0,-0.0488,0.0,3.0,-0.0488,0.0,3.0
|
| 43 |
zoo-mono_vs_joint,multiblimp:nld,16.0,16.0,0.7432,0.0,0.0,,0.0,0.0,-0.1536,0.0,16.0,-0.1525,2.0,16.0
|
| 44 |
zoo-mono_vs_joint,multiblimp:pol,3.0,3.0,0.768,0.0,0.0,,0.0,0.0,-0.0407,0.0,3.0,-0.0372,0.0,3.0
|
| 45 |
zoo-mono_vs_joint,multiblimp:rus,3.0,3.0,0.757,0.0,0.0,,0.0,0.0,-0.0522,0.0,3.0,-0.0456,0.0,3.0
|
| 46 |
zoo-mono_vs_joint,multiblimp:tur,3.0,3.0,0.7512,0.0,0.0,,0.0,0.0,-0.0431,0.0,3.0,-0.0425,0.0,3.0
|
| 47 |
-
zoo-mono_vs_joint,zhoblimp,6.0,6.0,0.
|
|
|
|
| 10 |
humanscale-pair,multiblimp:nld,24.0,1.0,0.6042,0.125,0.125,0.0081,4.0,24.0,-0.2842,0.0,24.0,-0.2838,0.0,24.0
|
| 11 |
humanscale-pair,zhoblimp,24.0,1.0,0.5832,0.0,0.0,0.002,3.0,24.0,-0.1633,0.0,24.0,-0.1142,3.0,24.0
|
| 12 |
zoo-curriculum,blimp_eng,49.0,49.0,0.5925,0.0,0.0,,0.0,0.0,-0.0137,5.0,49.0,,0.0,0.0
|
| 13 |
+
zoo-curriculum,blimp_nl,14.0,14.0,0.765,0.0,0.0,,0.0,0.0,-0.0158,0.0,14.0,,0.0,0.0
|
| 14 |
zoo-curriculum,multiblimp:deu,11.0,11.0,0.8856,0.0,0.0,,0.0,0.0,-0.0179,2.0,11.0,,0.0,0.0
|
| 15 |
zoo-curriculum,multiblimp:eng,49.0,49.0,0.7711,0.0,0.0,,0.0,0.0,-0.0194,4.0,49.0,,0.0,0.0
|
| 16 |
zoo-curriculum,multiblimp:eus,3.0,3.0,0.978,0.0,0.0,,0.0,0.0,0.0037,2.0,3.0,,0.0,0.0
|
| 17 |
zoo-curriculum,multiblimp:hin,3.0,3.0,0.8659,0.0,0.0,,0.0,0.0,-0.0514,0.0,3.0,,0.0,0.0
|
| 18 |
zoo-curriculum,multiblimp:ita,3.0,3.0,0.8142,0.0,0.0,,0.0,0.0,-0.0343,0.0,3.0,,0.0,0.0
|
| 19 |
+
zoo-curriculum,multiblimp:nld,14.0,14.0,0.8828,0.0,0.0,,0.0,0.0,-0.0211,0.0,14.0,,0.0,0.0
|
| 20 |
zoo-curriculum,multiblimp:pol,3.0,3.0,0.7879,0.0,0.0,,0.0,0.0,-0.0329,0.0,3.0,,0.0,0.0
|
| 21 |
+
zoo-curriculum,multiblimp:rus,3.0,3.0,0.765,0.0,0.0,,0.0,0.0,-0.0531,0.0,3.0,,0.0,0.0
|
| 22 |
zoo-curriculum,multiblimp:tur,3.0,3.0,0.7773,0.0,0.0,,0.0,0.0,-0.0318,0.0,3.0,,0.0,0.0
|
| 23 |
+
zoo-curriculum,zhoblimp,3.0,3.0,0.7221,0.0,0.0,,0.0,0.0,0.0122,3.0,3.0,,0.0,0.0
|
| 24 |
+
zoo-language,blimp_eng,692.0,111.0,0.5249,0.0549,0.3844,0.0003,69.0,373.0,-0.0166,146.0,692.0,,0.0,0.0
|
| 25 |
+
zoo-language,blimp_nl,108.0,16.0,0.5295,0.1852,0.3889,0.0056,16.0,66.0,-0.2203,0.0,108.0,,0.0,0.0
|
| 26 |
+
zoo-language,multiblimp:deu,110.0,16.0,0.7629,0.0,0.0,-0.0031,7.0,70.0,-0.1316,0.0,110.0,,0.0,0.0
|
| 27 |
+
zoo-language,multiblimp:ell,78.0,13.0,0.7641,0.0,0.0,0.012,12.0,39.0,-0.1392,0.0,78.0,,0.0,0.0
|
| 28 |
+
zoo-language,multiblimp:eng,692.0,111.0,0.6656,0.0014,0.0,0.0013,66.0,373.0,-0.0212,186.0,692.0,,0.0,0.0
|
| 29 |
+
zoo-language,multiblimp:fin,78.0,13.0,0.6749,0.0,0.0,0.0016,7.0,39.0,-0.164,0.0,78.0,,0.0,0.0
|
| 30 |
+
zoo-language,multiblimp:hin,78.0,13.0,0.6188,0.2051,0.141,-0.0002,7.0,39.0,-0.2554,0.0,78.0,,0.0,0.0
|
| 31 |
+
zoo-language,multiblimp:isl,78.0,13.0,0.7219,0.0,0.0,0.0006,8.0,39.0,-0.1985,0.0,78.0,,0.0,0.0
|
| 32 |
+
zoo-language,multiblimp:ita,78.0,13.0,0.5988,0.2051,0.2179,0.007,12.0,39.0,-0.2101,0.0,78.0,,0.0,0.0
|
| 33 |
+
zoo-language,multiblimp:nld,108.0,16.0,0.5814,0.2222,0.2037,0.005,12.0,66.0,-0.3018,0.0,108.0,,0.0,0.0
|
| 34 |
+
zoo-language,multiblimp:pol,79.0,13.0,0.6231,0.0,0.0,0.0049,12.0,41.0,-0.1651,0.0,79.0,,0.0,0.0
|
| 35 |
+
zoo-language,multiblimp:rus,78.0,13.0,0.5582,0.1026,0.141,0.0064,12.0,39.0,-0.2359,0.0,78.0,,0.0,0.0
|
| 36 |
+
zoo-language,multiblimp:spa,78.0,13.0,0.6818,0.0,0.0,0.0055,9.0,39.0,-0.2159,0.0,78.0,,0.0,0.0
|
| 37 |
+
zoo-language,multiblimp:tur,78.0,13.0,0.6505,0.0,0.0,0.0016,8.0,39.0,-0.1159,0.0,78.0,,0.0,0.0
|
| 38 |
+
zoo-language,zhoblimp,108.0,16.0,0.4125,0.1759,0.3889,-0.0012,9.0,66.0,-0.2613,0.0,108.0,,0.0,0.0
|
| 39 |
zoo-mono_vs_joint,blimp_eng,50.0,50.0,0.5602,0.02,0.22,,0.0,0.0,-0.0368,6.0,50.0,-0.0166,14.0,50.0
|
| 40 |
+
zoo-mono_vs_joint,blimp_nl,16.0,16.0,0.6651,0.375,0.375,,0.0,0.0,-0.1153,2.0,16.0,-0.1133,4.0,16.0
|
| 41 |
+
zoo-mono_vs_joint,multiblimp:deu,13.0,13.0,0.7724,0.0,0.0,,0.0,0.0,-0.1266,0.0,13.0,-0.1228,1.0,13.0
|
| 42 |
+
zoo-mono_vs_joint,multiblimp:eng,50.0,50.0,0.6825,0.0,0.0,,0.0,0.0,-0.0778,3.0,50.0,-0.0609,12.0,50.0
|
| 43 |
zoo-mono_vs_joint,multiblimp:hin,3.0,3.0,0.8123,0.0,0.0,,0.0,0.0,-0.091,0.0,3.0,-0.091,0.0,3.0
|
| 44 |
zoo-mono_vs_joint,multiblimp:ita,3.0,3.0,0.7892,0.0,0.0,,0.0,0.0,-0.0488,0.0,3.0,-0.0488,0.0,3.0
|
| 45 |
zoo-mono_vs_joint,multiblimp:nld,16.0,16.0,0.7432,0.0,0.0,,0.0,0.0,-0.1536,0.0,16.0,-0.1525,2.0,16.0
|
| 46 |
zoo-mono_vs_joint,multiblimp:pol,3.0,3.0,0.768,0.0,0.0,,0.0,0.0,-0.0407,0.0,3.0,-0.0372,0.0,3.0
|
| 47 |
zoo-mono_vs_joint,multiblimp:rus,3.0,3.0,0.757,0.0,0.0,,0.0,0.0,-0.0522,0.0,3.0,-0.0456,0.0,3.0
|
| 48 |
zoo-mono_vs_joint,multiblimp:tur,3.0,3.0,0.7512,0.0,0.0,,0.0,0.0,-0.0431,0.0,3.0,-0.0425,0.0,3.0
|
| 49 |
+
zoo-mono_vs_joint,zhoblimp,6.0,6.0,0.4881,0.0,0.0,,0.0,0.0,-0.2373,0.0,6.0,-0.2163,3.0,6.0
|
summary_by_operator_arm.csv
CHANGED
|
@@ -1,7 +1,7 @@
|
|
| 1 |
operator,arm,n_cells,n_merges,mean_score,frac_at_chance,frac_near_chance,mean_d_naive,n_beats_naive,n_naive_cmp,mean_d_better_parent,n_beats_better_parent,n_parent_cmp,mean_d_ceiling,n_beats_ceiling,n_ceiling_cmp
|
| 2 |
-
average,aligned,38.0,10.0,0.
|
| 3 |
-
average,naive,
|
| 4 |
-
average,transport,
|
| 5 |
csp,aligned,10.0,3.0,0.5297,0.0,0.2,0.0,0.0,10.0,-0.2338,0.0,10.0,-0.1787,0.0,10.0
|
| 6 |
csp,naive,10.0,3.0,0.5297,0.0,0.2,0.0,0.0,10.0,-0.2338,0.0,10.0,-0.1787,0.0,10.0
|
| 7 |
csp,transport,10.0,3.0,0.538,0.2,0.2,0.0083,5.0,10.0,-0.2255,0.0,10.0,-0.1705,0.0,10.0
|
|
@@ -14,15 +14,15 @@ opcm,transport,10.0,3.0,0.5694,0.1,0.2,0.0722,8.0,10.0,-0.1941,0.0,10.0,-0.1391,
|
|
| 14 |
ortho_merge,aligned,10.0,3.0,0.5627,0.0,0.1,0.0,0.0,10.0,-0.2008,0.0,10.0,-0.1458,0.0,10.0
|
| 15 |
ortho_merge,naive,10.0,3.0,0.5627,0.0,0.1,0.0,0.0,10.0,-0.2008,0.0,10.0,-0.1458,0.0,10.0
|
| 16 |
ortho_merge,transport,10.0,3.0,0.5618,0.0,0.1,-0.001,6.0,10.0,-0.2018,0.0,10.0,-0.1467,1.0,10.0
|
| 17 |
-
qmd_guided,aligned,
|
| 18 |
-
qmd_guided,naive,
|
| 19 |
-
qmd_guided,transport,
|
| 20 |
svc,aligned,10.0,3.0,0.4972,0.1,0.1,0.0,0.0,10.0,-0.2663,0.0,10.0,-0.2113,0.0,10.0
|
| 21 |
svc,naive,10.0,3.0,0.4972,0.1,0.1,0.0,0.0,10.0,-0.2663,0.0,10.0,-0.2113,0.0,10.0
|
| 22 |
svc,transport,10.0,3.0,0.5694,0.1,0.2,0.0722,8.0,10.0,-0.1941,0.0,10.0,-0.1391,0.0,10.0
|
| 23 |
task_arithmetic,aligned,14.0,4.0,0.6794,0.0714,0.1429,0.0,0.0,14.0,-0.0772,0.0,14.0,-0.0394,1.0,10.0
|
| 24 |
-
task_arithmetic,naive,
|
| 25 |
-
task_arithmetic,transport,
|
| 26 |
-
ties,aligned,
|
| 27 |
-
ties,naive,
|
| 28 |
-
ties,transport,
|
|
|
|
| 1 |
operator,arm,n_cells,n_merges,mean_score,frac_at_chance,frac_near_chance,mean_d_naive,n_beats_naive,n_naive_cmp,mean_d_better_parent,n_beats_better_parent,n_parent_cmp,mean_d_ceiling,n_beats_ceiling,n_ceiling_cmp
|
| 2 |
+
average,aligned,38.0,10.0,0.5921,0.1316,0.2105,-0.0005,1.0,38.0,-0.1545,1.0,38.0,-0.0692,0.0,10.0
|
| 3 |
+
average,naive,38.0,10.0,0.5926,0.1316,0.2105,0.0,0.0,38.0,-0.154,1.0,38.0,-0.0692,0.0,10.0
|
| 4 |
+
average,transport,413.0,114.0,0.6132,0.0533,0.1574,-0.0082,15.0,34.0,-0.0971,53.0,413.0,-0.081,0.0,10.0
|
| 5 |
csp,aligned,10.0,3.0,0.5297,0.0,0.2,0.0,0.0,10.0,-0.2338,0.0,10.0,-0.1787,0.0,10.0
|
| 6 |
csp,naive,10.0,3.0,0.5297,0.0,0.2,0.0,0.0,10.0,-0.2338,0.0,10.0,-0.1787,0.0,10.0
|
| 7 |
csp,transport,10.0,3.0,0.538,0.2,0.2,0.0083,5.0,10.0,-0.2255,0.0,10.0,-0.1705,0.0,10.0
|
|
|
|
| 14 |
ortho_merge,aligned,10.0,3.0,0.5627,0.0,0.1,0.0,0.0,10.0,-0.2008,0.0,10.0,-0.1458,0.0,10.0
|
| 15 |
ortho_merge,naive,10.0,3.0,0.5627,0.0,0.1,0.0,0.0,10.0,-0.2008,0.0,10.0,-0.1458,0.0,10.0
|
| 16 |
ortho_merge,transport,10.0,3.0,0.5618,0.0,0.1,-0.001,6.0,10.0,-0.2018,0.0,10.0,-0.1467,1.0,10.0
|
| 17 |
+
qmd_guided,aligned,252.0,77.0,0.7049,0.0079,0.0238,0.0001,3.0,24.0,-0.0435,19.0,252.0,-0.0653,2.0,70.0
|
| 18 |
+
qmd_guided,naive,28.0,7.0,0.6067,0.0357,0.1071,0.0,0.0,28.0,-0.1311,4.0,28.0,,0.0,0.0
|
| 19 |
+
qmd_guided,transport,499.0,139.0,0.6332,0.0521,0.1523,-0.0136,7.0,28.0,-0.0837,62.0,499.0,-0.0755,34.0,96.0
|
| 20 |
svc,aligned,10.0,3.0,0.4972,0.1,0.1,0.0,0.0,10.0,-0.2663,0.0,10.0,-0.2113,0.0,10.0
|
| 21 |
svc,naive,10.0,3.0,0.4972,0.1,0.1,0.0,0.0,10.0,-0.2663,0.0,10.0,-0.2113,0.0,10.0
|
| 22 |
svc,transport,10.0,3.0,0.5694,0.1,0.2,0.0722,8.0,10.0,-0.1941,0.0,10.0,-0.1391,0.0,10.0
|
| 23 |
task_arithmetic,aligned,14.0,4.0,0.6794,0.0714,0.1429,0.0,0.0,14.0,-0.0772,0.0,14.0,-0.0394,1.0,10.0
|
| 24 |
+
task_arithmetic,naive,17.0,5.0,0.6791,0.0588,0.1176,0.0,0.0,17.0,-0.0675,0.0,17.0,-0.0394,1.0,10.0
|
| 25 |
+
task_arithmetic,transport,413.0,114.0,0.6367,0.0581,0.1477,-0.0025,5.0,13.0,-0.0736,67.0,413.0,-0.0443,1.0,10.0
|
| 26 |
+
ties,aligned,417.0,115.0,0.5896,0.0624,0.1727,0.0,2.0,417.0,-0.121,50.0,417.0,-0.1334,0.0,10.0
|
| 27 |
+
ties,naive,417.0,115.0,0.5896,0.0624,0.1727,0.0,0.0,417.0,-0.121,50.0,417.0,-0.1334,0.0,10.0
|
| 28 |
+
ties,transport,413.0,114.0,0.5965,0.0436,0.1719,0.0075,244.0,413.0,-0.1138,52.0,413.0,-0.1327,0.0,10.0
|