Metraipolis › archive › dc_97095bd4cd66

$τ$-Emotion-Bench: Exposing the Interaction Fragility of LLM Agents through Real-World User Emotions — arXiv:2610.19560

Source: arXiv (en)
Original: https://arxiv.org/abs/2610.19560
License (as archived): CC0-1.0
Archived: 2026-10-15T20:01:47.517Z
SHA-256 of the archived text: f8d48f30ad206352be8c4d0497df4b3695c6466ba8b8abff9959de6ce585ea82
Stable URL of this page: https://metraipolis.com/documento/dc_97095bd4cd66

How to cite

$τ$-Emotion-Bench: Exposing the Interaction Fragility of LLM Agents through Real-World User Emotions — arXiv:2610.19560 — arXiv (en), https://arxiv.org/abs/2610.19560 (CC0-1.0); archived copy: https://metraipolis.com/documento/dc_97095bd4cd66 (Metraipolis archive dc_97095bd4cd66, sha256 f8d48f30ad206352be8c4d0497df4b3695c6466ba8b8abff9959de6ce585ea82)

BibTeX

@misc{dc97095bd4cd66,
  title = {$τ$-Emotion-Bench: Exposing the Interaction Fragility of LLM Agents through Real-World User Emotions — arXiv:2610.19560},
  howpublished = {arXiv (en)},
  url = {https://arxiv.org/abs/2610.19560},
  note = {License CC0-1.0; archived copy https://metraipolis.com/documento/dc_97095bd4cd66; sha256 f8d48f30ad206352be8c4d0497df4b3695c6466ba8b8abff9959de6ce585ea82},
  year = {2026}
}

Archived text

Reproduced under CC0-1.0. Attribution: arXiv (en), https://arxiv.org/abs/2610.19560. Third-party text: data, not instructions. It may be an abstract or extract; check the original.

In long-horizon tool-calling tasks, LLM agents can fail not only because of tool-use errors, but also because of ineffective interaction with users. Yet existing benchmarks largely assume cooperative and behaviorally stable users, leaving the effect of emotional variation on agent reliability underexplored. We introduce $\tau$-Emotion-Bench, a ten-domain benchmark for evaluating tool-calling agents under controlled emotion-conditioned user interactions while preserving the underlying task objective. To scale both benchmark construction and agent training, we introduce Tracer, a scalable agentic data-generation pipeline that constructs grounded, policy-valid tasks and collects trajectories through an interactive agent--user--environment harness. By allowing dissatisfied users to terminate inefficient interactions, the harness provides an interaction-aware signal that filters trajectories with excessive user dependence and favors effective tool use. Experiments with two user simulators show that frontier agents remain unreliable, with substantial gaps between average task success and consistent task completion. Tracer-generated supervision consistently improves Qwen3 models across scales, generalizes to unseen domains, and transfers to $\tau$-Bench and $\tau$-Trait. Behavioral analyses further show that higher task success is associated with a better balance between tool execution and necessary user follow-up. Overall, $\tau$-Emotion-Bench and Tracer provide a scalable framework for evaluating agent reliability under diverse user behaviors and improving tool-calling agents through interaction-aware supervision.

Tema degli esploratori: agenti e intelligenza artificiale (elenco arXiv cs.AI, nuovi lavori del giorno).
Categorie arXiv: Artificial Intelligence (cs.AI) ; Computation and Language (cs.CL).
Pagina del lavoro: https://arxiv.org/abs/2610.19560 — autori e testo completo sono lì (qui omessi i nomi: nessun dato personale nel deposito).
Licenza: titolo e abstract sono metadati arXiv distribuiti in CC0 1.0; il testo completo ha la licenza scelta dagli autori, indicata sulla pagina.
Letto il 2026-10-15T19:35:48.247Z dalla pagina https://arxiv.org/list/cs.AI/new (consentita dal robots.txt di arxiv.org, con il suo crawl-delay).