Update README.md
Browse files
README.md
CHANGED
|
@@ -5,6 +5,31 @@ license: apache-2.0
|
|
| 5 |
|
| 6 |
WIP
|
| 7 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 8 |
## GPQA
|
| 9 |
|
| 10 |
|
|
|
|
| 5 |
|
| 6 |
WIP
|
| 7 |
|
| 8 |
+
|
| 9 |
+
| Tasks |**HuggingFaceTB/SmolLM2-360M** Value|**aloobun/d-SmolLM2-360M** Value|
|
| 10 |
+
|----------------------------------------------------------|-------------:|-------------:|
|
| 11 |
+
| - leaderboard_bbh_causal_judgement | 0.4545 | 0.4652 |
|
| 12 |
+
| - leaderboard_bbh_geometric_shapes | 0.1680 | 0.2040 |
|
| 13 |
+
| - leaderboard_bbh_movie_recommendation | 0.2120 | 0.2440 |
|
| 14 |
+
| - leaderboard_bbh_penguins_in_a_table | 0.2055 | 0.2123 |
|
| 15 |
+
| - leaderboard_bbh_reasoning_about_colored_objects | 0.1160 | 0.1320 |
|
| 16 |
+
| - leaderboard_bbh_ruin_names | 0.2360 | 0.2480 |
|
| 17 |
+
| - leaderboard_bbh_salient_translation_error_detection | 0.1480 | 0.2120 |
|
| 18 |
+
| - leaderboard_bbh_snarks | 0.5169 | 0.5281 |
|
| 19 |
+
| - leaderboard_bbh_temporal_sequences | 0.2720 | 0.2800 |
|
| 20 |
+
| - leaderboard_musr_murder_mysteries | 0.5040 | 0.5160 |
|
| 21 |
+
|
| 22 |
+
|
| 23 |
+
|
| 24 |
+
# Eval Results aloobun/d-SmolLM2-360M
|
| 25 |
+
|
| 26 |
+
Todo:
|
| 27 |
+
|
| 28 |
+
ifeval (0-shot, generative)
|
| 29 |
+
|
| 30 |
+
Math-lvl-5 (4-shots, generative, minerva version)
|
| 31 |
+
|
| 32 |
+
|
| 33 |
## GPQA
|
| 34 |
|
| 35 |
|