τ-Emotion-Bench exposes LLM agent fragility under emotional users
A paper accepted to the UserSim Workshop at NeurIPS 2026 introduces τ-Emotion-Bench, a ten-domain benchmark that evaluates tool-calling agents under controlled emotion-conditioned user interactions while preserving the underlying task objective. The authors also present Tracer, a scalable agentic data-generation pipeline that builds policy-valid tasks and collects trajectories, letting dissatisfied users terminate inefficient interactions so trajectories with excessive user dependence are filtered out. Experiments with two user simulators show frontier agents remain unreliable, with substantial gaps between average task success and consistent task completion, while Tracer-generated supervision improves Qwen3 models across scales and transfers to τ-Bench and τ-Trait.
User emotion is a common but often ignored variable in agent evaluation, and this work offers a scalable benchmark plus an interaction-aware training signal.