Weak-Driven Learning: How Weak Agents make Strong Agents Stronger
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2026
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908822317039616 |
|---|---|
| author | Chen, Zehao Li, Gongxun Ai, Tianxiang Li, Yifei Huang, Zixuan Zhou, Wang Zhuang, Fuzhen Liu, Xianglong Li, Jianxin Wang, Deqing Ban, Yikun |
| author_facet | Chen, Zehao Li, Gongxun Ai, Tianxiang Li, Yifei Huang, Zixuan Zhou, Wang Zhuang, Fuzhen Liu, Xianglong Li, Jianxin Wang, Deqing Ban, Yikun |
| contents | As post-training optimization becomes central to improving large language models, we observe a persistent saturation bottleneck: once models grow highly confident, further training yields diminishing returns. While existing methods continue to reinforce target predictions, we find that informative supervision signals remain latent in models' own historical weak states. Motivated by this observation, we propose WMSS (Weak Agents Can Make Strong Agents Stronger), a post-training paradigm that leverages weak checkpoints to guide continued optimization. By identifying recoverable learning gaps via entropy dynamics and reinforcing them through compensatory learning, WMSS enables strong agents to improve beyond conventional post-training saturation. Experiments on mathematical reasoning and code generation datasets show that agents trained with our approach achieve effective performance improvements, while incurring zero additional inference cost. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2602_08222 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | Weak-Driven Learning: How Weak Agents make Strong Agents Stronger Chen, Zehao Li, Gongxun Ai, Tianxiang Li, Yifei Huang, Zixuan Zhou, Wang Zhuang, Fuzhen Liu, Xianglong Li, Jianxin Wang, Deqing Ban, Yikun Artificial Intelligence As post-training optimization becomes central to improving large language models, we observe a persistent saturation bottleneck: once models grow highly confident, further training yields diminishing returns. While existing methods continue to reinforce target predictions, we find that informative supervision signals remain latent in models' own historical weak states. Motivated by this observation, we propose WMSS (Weak Agents Can Make Strong Agents Stronger), a post-training paradigm that leverages weak checkpoints to guide continued optimization. By identifying recoverable learning gaps via entropy dynamics and reinforcing them through compensatory learning, WMSS enables strong agents to improve beyond conventional post-training saturation. Experiments on mathematical reasoning and code generation datasets show that agents trained with our approach achieve effective performance improvements, while incurring zero additional inference cost. |
| title | Weak-Driven Learning: How Weak Agents make Strong Agents Stronger |
| topic | Artificial Intelligence |
| url | https://arxiv.org/abs/2602.08222 |