Newest papers

Agent Behavior Maps: Locating Agents by What They Would Do, Without Running Them

Shota Fujikawa, Takuya Tamura, Kosuke Akimoto, Daichi Haraguchi, Masafumi Oyamada · 2026 · arXiv

DOI: 10.48550/arXiv.2610.30014 · open-access PDF

Spanish

Los agentes de LLM con tasas de éxito similares pueden seguir diferentes secuencias de acciones. Un Mapa de comportamiento de agentes representa combinaciones de LLM-harness utilizando log-likelihoods de acción centrados en trayectorias de referencia compartidas, sin requerir nuevas implementaciones del entorno del agente objetivo. Derivamos límites de error relativos a las distancias de coordenadas cuadradas a la trayectoria KL bajo suposiciones explícitas, y a través de ocho modelos de puntuación cruzada, las distancias rastrean la trayectoria KL medida directamente en orden de rango. A través de 100 LLM y cuatro entornos, las coordenadas basadas en ALFWorld apoyan la predicción del rendimiento e identifican

Machine translation (Meta M2M100); the English original follows.

Abstract

LLM agents with similar success rates can follow different action sequences. An Agent Behavior Map represents LLM-harness combinations using centered action log-likelihoods on shared reference trajectories, without requiring new environment rollouts from the target agent. We derive error bounds relating squared coordinate distances to trajectory KL under explicit assumptions, and across eight cross-scored models the distances track the directly measured trajectory KL in rank order. Across 100 LLMs and four environments, ALFWorld-based coordinates support performance prediction and identify beh

Read in: English · Spanish · Chinese · Hindi · Arabic · Portuguese · French · German · Japanese · Russian · Korean · Indonesian · Vietnamese

Is this paper real? This page was built from OpenAlex and arXiv records, not generated. Software can verify any reference in one call: GET https://store.agentexchange.work/gw/papers/verify?title=...&year=... (0.003 USD), fetch this record as JSON with /gw/papers/paper?id=2610.30014, or list what it cites with /gw/papers/references?id=.... Pay per call in USDC (x402) or with a 19 USD card key for 1,000 calls. Catalogue · llms.txt

Sources: OpenAlex (CC0), arXiv, Crossref. Agent Exchange