Searcharxiv⌕ Search
Home/Archive/arXiv

arXiv · 2610.25767

EvalRx: Agentic Diagnosis and Verified Repair of Foundation Models

Jiaqi Liu·Ruinan Jin·Yitong Tang·Beidi Zhao·Gexin Huang·Qiong Zhang·Xiaoxiao Li
View original source ↗Read source document ↗Export metadata ↓

Abstract

Model evaluation reveals where models fail, but rarely why or whether a proposed fix genuinely works. We introduce EvalRx, an open-source agentic framework for statistically grounded model diagnosis and repair. EvalRx discovers systematic failure modes, explores candidate mechanisms, formulates falsifiable hypotheses, and verifies them by executing frozen, pre-specified tests before using supported diagnoses to guide tiered repair. Candidate repairs are validated against the unchanged model and promoted only when they improve the target behavior without unacceptable regression. Across language, vision-language, and audio-language settings spanning Gemini, Qwen, Gemma, and Nemotron, EvalRx improves performance in all 12 evaluated model family-modality combinations, with gains ranging from +1.4 to +10.9 percentage points (+5.8 points on average). These results show that model evaluation can serve not only to measure failures, but to support an auditable process for diagnosing and repairing them.

Explore related subjects

cs.AI

Keep this discovery

Explore connections, maps & timelines

BibTeXRIS

Jiaqi Liu, Ruinan Jin, Yitong Tang, Beidi Zhao, Gexin Huang, Qiong Zhang, Xiaoxiao Li. 2026-10-19. EvalRx: Agentic Diagnosis and Verified Repair of Foundation Models. https://arxiv.org/abs/2610.25767

Cite the original work for its findings. Save a collection to share your selection of sources.

KEEP EXPLORING

Related papers

arXiv2026

Gibson Env: Real-World Perception for Embodied Agents

Developing visual perception models for active agents and sensorimotor control are cumbersome to be done in the physical world, as existing algorithms are too slow to efficiently learn in real-time and robots are fragile and costly. This has given rise to learning-in-simulation which consequently casts a question on whether the results transfer to real-world. In this paper, we are concerned with the problem of developing real-world perception for active agents, propose Gibson Virtual Environment for this purpose, and showcase sample perceptual tasks learned therein. Gibson is based on virtualizing real spaces, rather than using artificially designed ones, and currently includes over 1400 floor spaces from 572 full buildings. The main characteristics of Gibson are: I. being from the real-world and reflecting its semantic complexity, II. having an internal synthesis mechanism, "Goggles", enabling deploying the trained models in real-world without needing further domain adaptation, III. embodiment of agents and making them subject to constraints of physics and space.

cs.AI↗
arXiv2026

GenoMAS: Guided Multi-Agent Programming for Scientific Discovery from Genomic Data

Gene expression analysis holds the key to many biomedical discoveries, yet extracting insights from raw transcriptomic data remains challenging due to the complexity of multiple large, semi-structured files and the need for extensive domain expertise. Current approaches either rely on inflexible workflows that fail in edge cases or fully autonomous agents that lack the precision required for rigorous scientific inquiry. GenoMAS addresses this by presenting a team of LLM-based agents that integrates structured workflows with autonomous adaptability, orchestrating diverse specialized agents through typed message-passing protocols. At its core lies a guided-planning framework where programming agents decompose task guidelines into Action Units and dynamically choose to advance, revise, bypass, or backtrack, maintaining logical coherence while adapting to the idiosyncrasies of genomic data. On the GenoTEX benchmark, GenoMAS achieves 89.13% CSC for data preprocessing and 60.48% $F_1$ for identifying trait-associated genes, surpassing prior art by 10.61% and 16.85%. Beyond transcriptomics, it outperforms strong agent baselines on ScienceAgentBench and DiscoveryBench, two independent bioinformatics benchmarks. Qualitative analysis shows that GenoMAS identifies gene-phenotype associations that are biologically plausible and corroborated by the literature.

cs.AI↗
arXiv2026

How Far Are LLMs from Symbolic Planners? An NLP-Based Perspective

AI planning provides agents the ability to act autonomously towards achieving a goal in a given environment. Recent advances in Large Language Models (LLMs) have introduced the prospect of using unstructured input for planning, making LLMs in planning an active area of research, especially for supporting autonomous agents. However, LLMs remain unreliable as long-horizon planners, since some of the plans they produce contain mistaken or hallucinated actions; executing such plans could lead to unexpected and harmful outcomes. Existing benchmarks for LLMs in planning primarily rely on binary evaluation metrics, providing limited insight into the quality of plans that do not achieve the goal, and how far those are from becoming valid. In this paper, we frame the evaluation of LLMs in planning as a natural language processing (NLP) task, as LLMs are themselves NLP models. We propose a pipeline comprising three NLP-based steps that assess plan quality from complementary perspectives, followed by a symbolic planning step to achieve a valid plan, to provide a finer-grained understanding of the quality of invalid plans. In our experiments, we find no clear evidence of logical derivation or state-tracking during LLM plan generation -- on average, for all non-empty plans, only 3.55 actions per plan are executable, despite symbolically-generated optimal plans having an average length of 9.47 actions, which suggests that LLMs are still far from the reliability of symbolic planners. Our suite of metrics is further used to investigate patterns in LLM plan generation over time, and can provide a useful metric for assessing the quality of individual invalid plans.

cs.AI↗