The authors built SkillTransferBench, comprising 35 task families and 408 executable cases across six domains.
Interleaved teaching helps agents learn skills that transfer better
Socratic Skill Learning improved reuse on held-out tasks while issue-focused review sharply reduced the expert supervision required.
Chinese Tech
Letian Zhu · Han Li · Yifan Yao · Xinping Lei · Xinyu Che · Jiahao Wang · +6 more
Nanjing University · Alibaba Group
Research Digest··3 min read
Zhu et al.
Why this paper
From Alibaba Group and Nanjing University
In one line
Interleaving expert teaching with student practice improves transferable skill learning for LLM agents.
What we could check
- ·No code link found
- ·No weights link found
- ·No dataset link found
- ·No compute details found
- ✓Limitations stated by the authors
- ✓Reports numbers on named benchmarks
Observed from the paper text and links we have. Absence here means we did not find it, not that it does not exist.
§