IncorrectTraceSFT Collection Math models and data for How Should Incorrect Traces Be Used in Supervised Fine-Tuning? • 16 items • Updated 17 days ago
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-correct-only Text Generation • 8B • Updated 17 days ago • 242
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-random-mask-90 Text Generation • 8B • Updated 17 days ago • 249
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-random-mask-80 Text Generation • 8B • Updated 17 days ago • 235
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-correct-only Text Generation • 8B • Updated 17 days ago • 242
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-random-mask-70 Text Generation • 8B • Updated 17 days ago • 233
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-random-mask-90 Text Generation • 8B • Updated 17 days ago • 249
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-random-mask-80 Text Generation • 8B • Updated 17 days ago • 235
suryadv/IncorrectTraceSFT-Qwen2.5-7B-MATH-random-mask-50 Text Generation • 8B • Updated 17 days ago • 233