developer tools
Renaming a Tool Should Not Erase the Skill
Cross-Interface Policy Regularization shows the same training task under two action vocabularies and penalizes behavioral disagreement between them.
Summary
Cross-Interface Policy Regularization shows the same training task under two action vocabularies and penalizes behavioral disagreement between them.
An agent can learn a task and still fail when an equivalent tool gets a new name or notation. CIPR replaces action names with uninformative identifiers and pairs each training example across two known-equivalent interfaces. A consistency loss aligns the resulting action distributions. On held-out function-calling perturbations, ordinary post-training loses nearly 25 percent accuracy; the authors report recovering 10.5 percent with CIPR while preserving performance on the original interface. The method does not solve unseen tool semantics, but it directly targets brittle dependence on surface labels.
Why it matters
Cross-Interface Policy Regularization shows the same training task under two action vocabularies and penalizes behavioral disagreement between them.
Limits and context
- The method does not solve unseen tool semantics, but it directly targets brittle dependence on surface labels.
Key claims
Cross-Interface Policy Regularization shows the same training task under two action vocabularies and penalizes behavioral disagreement between them.
Qualification: The method does not solve unseen tool semantics, but it directly targets brittle dependence on surface labels.
Evidence: source-2026-10-15-013
Sources
- arXiv preprint 2610.20576, original v1arXiv / submitting research authors · primary research
Corrections
: Publication note: This October 15 edition was prepared before its scheduled 9:00 AM America/Detroit reveal, but publication was delayed by a production build failure. The original reporting and image-generation timestamps are preserved.