4 days ago
-
Update SenseMath dataset and runnable evaluation entry points
-
Align Judge data with paper J1 and J2; remove J3
Keep Apply unchanged; retain J1 full 300 and separate filtered 261; replace error detection with J2 strategy identification 80.
-
Upload data/judge_j3.json with huggingface_hub
-
Upload data/judge_j2.json with huggingface_hub
-
Upload data/judge_j1.json with huggingface_hub
-
Upload data/sensemath_v2_d16.json with huggingface_hub
-
Upload data/sensemath_v2_d8.json with huggingface_hub
-
Upload data/sensemath_v2_d4.json with huggingface_hub
-
Upload data/sensemath_v2_d2.json with huggingface_hub
