arXiv：cs.AI· Jiandong Ding, Honglei Ji, Ming Liu, Tao Duan·· 3 小时前
SkillContrast：面向 Agent 技能重排的差异引导文本选择
SkillContrast: Difference-Guided Text Selection for Agent Skill Reranking
AI 导读
SkillContrast 是一种免训练选择器，通过对比检索到的技能并只保留差异文本及局部上下文，供预训练重排器使用。在 SameCapRisk-Bench 的 1235 条请求上，相同候选输入长度下，它比 TF-IDF 查询选择多出 54-72 个 clean hits，覆盖 2 种检索器与 2 种重排器规模。
正文
Abstract:Similar agent skills can share instructions but differ in their conditions of use. Query-based text selection may retain shared instructions and omit these distinctions. We introduce SkillContrast, a training-free selector that compares retrieved skills and retains their differing text with local context for a pretrained reranker. On 1,235 requests from SameCapRisk-Bench, it yields 54-72 more clean hits (requests that retrieve a helpful skill without its marked risky sibling) than TF-IDF query selection at identical per-candidate input lengths, across 2 retrievers and 2 reranker sizes. Length-matched component replacements identify differing text as the main contributor in the primary setting, with smaller, mixed context effects. Relative to full skill bodies, SkillContrast uses 51.1-58.8% fewer model-input tokens, with 10-18 fewer clean hits at 0.6B and matching or higher observed clean-hit counts at 4B. Candidate-relative differences thus complement query relevance in selecting compact reranking inputs.
| Comments: | 5 pages, 2 figures, 3 tables |
| Subjects: | Information Retrieval (cs.IR); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.11650 [cs.IR] |
| (or arXiv:2610.11650v1 [cs.IR] for this version) | |
| https://doi.org/10.48550/arXiv.2610.11650 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jiandong Ding [view email]
[v1]
Thu, 8 Oct 2026 10:25:33 UTC (2,575 KB)
来源：arXiv：cs.AI · arxiv.org