Holistic Evaluation and Failure Diagnosis of AI Agents
Netta Madvil, Gilad Dym, Alon Mecilati, Edo Dekel, · 2026
Read Holistic Evaluation and Failure Diagnosis of AI Agents by Netta Madvil, Gilad Dym, Alon Mecilati, Edo Dekel, (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-14865.txt.
Description
Madvil and team present a holistic framework for evaluating AI agents and their failures. Outcome metrics report success or failure with no explanation why, and process-level approaches struggle to tie failure types to precise locations in long traces. The framework pairs top-down diagnosis with bottom-up span-level scoring, decomposing analysis into independent per-span assessments with rationales. On TRAIL, state-of-the-art results are claimed on GAIA and SWE-Bench, with gains up to thirty-eight percent on category F1 and severalfold gains on localisation.
Catalog record
| CREATOR | Netta Madvil, Gilad Dym, Alon Mecilati, Edo Dekel, |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 49150 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2605.14865