AdaMAST: Adaptive Failure Taxonomies for Improving LLM Agents
AdaMAST introduces an adaptive failure taxonomy that automatically derives named failure modes from an LLM agent’s execution traces. The taxonomy is organized along three fixed axes—system-level, role-specific, and domain-specific—and is validated before use in improvement loops such as best‑of‑N judging, runtime feedback, and mutation feedback. Experiments show notable performance gains across several benchmarks without requiring hand‑crafted codes or human annotation.
- ▪AdaMAST creates a bespoke taxonomy by inducing failure codes from the agent’s own rollouts, eliminating the need for pre‑written rubrics.
- ▪The adaptive taxonomy is applied in three improvement loops, including best‑of‑N selection, runtime feedback for coding agents, and mutation feedback in search optimization.
- ▪Across benchmarks, AdaMAST improves success rates by up to 15 points over fixed taxonomies and achieves up to 91.9% accuracy on held‑out problems.
- ▪The system operates fully automatically, using an agreement gate to validate each generated taxonomy before deployment.
Hacker News (AI / LLM) files mainly under ai. We currently carry 3,007 of its stories.
Story provenance
Source · retrieval · rights · ranking — open for full record
inspect →
Story provenance
Attribution is not the same as permission. This drawer separates discovery metadata, excerpts, WeSearch-generated summaries, reuse status, and whether the publisher receives the visit. Nothing here claims a legal grant the publisher has not made.
Record
| Original publisher | Github |
| Canonical URL | https://multi-agent-systems-failure-taxonomy.github.io/AdaMAST/blogs/adamast_paper/ |
| Publication time | Fri, 31 Jul 2026 06:22:31 +0000 |
| Retrieval time | 2026-07-31T06:42:50.389Z |
| Last seen | 2026-07-31T06:42:50.389Z |
| Headline source | Publisher (no WeSearch rewrite) |
| Excerpt source | publisher body |
| Excerpt method | First ~120 words (~800 chars) of extracted publisher body, fair-use limited. |
| Summary | WeSearch · cerebras-chat (WeSearch summarizer) |
| Summary source text | contentText |
| Citation coverage | Summary is a WeSearch-generated derivative; primary citation is the original publisher URL. |
| Cluster | IupwGerunp1s · 1 stories |
| Cluster logic | Grouped by semantic title/content similarity across sources within a rolling window. Same-publisher template collisions are excluded from coverage comparison. |
| Ranking reason | Story pages are not engagement-ranked. Hub feeds use recency, with optional source-diversified chronological ordering (cap consecutive stories per source). No personalized ranking. |
| Publisher visit | Yes — open original |
| Substitutes article? | No — link-out required for full text |
Rights status (four layers)
WeSearch handling by dimension
| Indexing | May the item be indexed (stored, ranked, made findable)? | Allowed |
| Snippet | May a short excerpt of the publisher's text be shown? | Allowed |
| AI summary | May WeSearch generate its own short summary of the article? | Limited |
| Retrieval / RAG | May the content be exposed for third-party retrieval-augmented generation? | Not asserted |
| Model training | May the content be used to train AI models? | Not asserted |
| Commercial reuse | May the content be reused commercially? | Not permitted |
Basis: Derived from the published RSS/Atom feed. Contact: [email protected]. Reviewed: 2026-07-24.
Opening excerpt (first ~120 words) tap to expand
Research blog AdaMAST: Adaptive Failure Taxonomies for Improving LLM Agents Fantastic Adaptive Taxonomies and How to Use Them Mert Cemri1*, Andrei Cojocaru1*, Melissa Pan1, Shu Liu1, Shubham Agarwal1, Alexander Krentsel1, Jay Tang2, Kannan Ramchandran1, Joseph E. Gonzalez1, Matei Zaharia1, Alexandros G. Dimakis1,3, Ion Stoica1 1 UC Berkeley 2 Apple 3 Bespoke Labs · * equal contribution July 2026 🤖 A failure taxonomy is a rubric of an agent system's recurring failure modes. Each mode gets a name, a definition, and evidence quoted from the system's own rollouts. Fixed, hand-built taxonomies such as MAST already work well as debugging aids. But a rubric written before a system exists cannot name the failures specific to that system's roles, harness, or task.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at Github.
Discussion
0 commentsMore from Github
