Conformal Agent Error Attribution
About
When multi-agent systems (MAS) fail, identifying where the decisive error occurred is the first step for automated recovery to an earlier state. Error attribution remains a fundamental challenge due to the long interaction traces that large language model-based MAS generate. This paper presents a framework for error attribution based on conformal prediction (CP) which provides finite-sample, distribution-free coverage guarantees. We introduce new algorithms for filtration-based CP designed for sequential data such as agent trajectories. Unlike existing CP algorithms, our approach predicts sets that are contiguous sequences to enable efficient recovery and debugging. We verify our theoretical guarantees on a variety of agents and datasets, show that errors can be precisely isolated, then use prediction sets to rollback MAS to correct their own errors. Our overall approach is model-agnostic, and offers a principled uncertainty layer for MAS error attribution. We release code at https://github.com/layer6ai-labs/conformal-agent-error-attribution.
Related benchmarks
| Task | Dataset | Result | Rank | |
|---|---|---|---|---|
| Error Attribution | GSM8k Right Dense | Removal Rate64 | 30 | |
| Error Attribution | GSM8k Mid Dense | Removal Rate63 | 30 | |
| Error Attribution | GSM8k Left Dense | Removal Rate71 | 30 | |
| Error Attribution | MATH | Removal Rate29 | 30 | |
| Error Attribution | Who&When | -- | 30 | |
| Step-level error discrimination | MATH and GSM8k (test) | -- | 4 | |
| Automated agent rollback | GSM8K Left Dense variant (test) | Success Rate77 | 3 | |
| Automated agent rollback | GSM8K Mid Dense variant (test) | Success Rate68 | 3 | |
| Automated agent rollback | GSM8K Right Dense variant (test) | Success Rate75 | 3 |