arXiv 2610.29680 · cs.AI · 2026-10-21
TraceForge: Distilling with Your Own Thinking
Yan Wang
Independent Research
在最新榜 #318 本站收录于 2026-10-24
一句话要点
TraceForge 讨论一些前沿模型只返回答案或简短摘要、不给完整推理过程时，怎样用自己的模型生成推理链来做蒸馏：作者认为靠跨模型重放加密推理块来恢复隐藏推理违反服务条款，不能长期依赖，提出“用自己的思考来蒸馏”的替代路线。
TraceForge addresses distillation when frontier APIs return only answers or short summaries without reasoning traces: rather than recovering hidden reasoning by replaying encrypted blocks across models, which breaks providers' terms of service, it proposes distilling with reasoning traces generated by your own model.
阅读 arXiv 原文 PDF Hugging Face 讨论 项目页
主题与学科
同主题：推理
- LEV: Thinking Fast and Deep with Looped Decision Models ▲337
- Thinking Inertia: LLMs Keep Thinking When Told Not To ▲241
- Rethinking Latent Visual Reasoning: Grounding Latent Reasoning in Visual Evidence ▲224
- Questioning the Questions: Sustaining Self-Evolution in Reasoning Models ▲85
- Improving Test-Time Scaling with Adaptive Looped Transformers ▲62
同主题：大语言模型
- Jeva: System One Models from First Principles 2026-10-13 · ▲371
- Post-Training Leaves Behavioral Shadows on Unrelated Decisions 2026-09-24 · ▲273
- WeVisOne-VL: Unified Visual Perception from Objects to Pixels 2026-10-22 · ▲247
- OmniEdu: Open Foundation Models for Learning and Teaching 2026-09-19 · ▲240
- SpatialScout: Can VLM Agents Actively Explore and Map 3D Scenes? 2026-10-19 · ▲170
数据说明
元数据来自 arXiv API（已核实）；点赞来自 Hugging Face 每日论文；引用来自 Semantic Scholar。最近刷新 2026-10-26。
要点由本站根据摘要自写，不转载摘要原文，请阅读原文。