SchemaShield-Lite：无需重训即可修复 LLM 智能体的接口漂移问题
Robust Function Calling for LLM Agents Under Interface Drift
做智能体工具调用的朋友看这篇，工具接口一改模型就崩，SchemaShield-Lite 不用重训就能在 BFCL-200 上把准确率从 0.58 拉到 0.805。
arXiv 论文研究了工具调用中的接口漂移：工具描述、参数 schema 和候选工具列表在部署后变化，导致模型调用失败。作者提出的 SchemaShield-Lite 是一个免训练的修复层，结合 schema 校验、候选集重试和单次受限修复。在 Qwen3.5-9B 上，该方法将 DICE-300 准确率从 0.717 提升到 0.813，BFCL-200 准确率从 0.580 提升到 0.805，分别挽回 22 个漂移失败中的 19 个和 45 个中的 33 个。消融实验显示候选集漂移（尤其是干扰工具）是主要损伤来源，单纯的重试策略无法解释这些收益。
Robust Function Calling for LLM Agents Under Interface Drift
Tool-augmented language-model agents rely on interface text: tool descriptions, argument schemas, and the candidate tool list shown at inference time. These interfaces often change after deployment even when the underlying tool semantics remain fixed, creating a practical reliability problem for function-calling systems. We study this failure mode as interface drift and evaluate whether a lightweight, training-free repair layer can recover robustness without retraining the base model. Our method, SchemaShield-Lite, combines schema validation, candidate-set retry for missing tool names, and one-shot constrained repair. On Qwen3.5-9B, it raises DICE-300 accuracy from 0.717 to 0.813 and BFCL-200 accuracy from 0.580 to 0.805, recovering 19/22 and 33/45 drift-induced failures with no observed repair harms. Naive retry does not explain the gains. Ablations show that candidate-set drift, especially distractor tools, accounts for most observed damage, while description and schema edits alone are less harmful. The results position validation-and-repair as a practical robustness layer for deployed LLM-agent tool use.
- DavidSacks10-14 15:45原文
