$τ$-Rec: A Verifiable Benchmark for Agentic Recommender Systems
As recommender systems transition toward agentic, multi-turn conversational interfaces, evaluation paradigms have struggled to keep pace. Current benchmarks often rely on "LLM-as-a-judge" evaluations, which introduce subjectivity, high costs and inconsistency. We present $τ$-Rec, a benchmark for agentic recommender systems that replaces subjective evaluation with verifiable rewards and a reveal-tagged elicitation (RTE) mechanism that controls how task constraints surface during dialogue. By testing agents against structured catalog predicates and employing a pass^k reliability metric, $τ$-Rec provides a systematic test for consistent reasoning. Our evaluation of nine configurations across...
| Authors | Bharath Sivaram Narasimhan, Karthik R Narasimhan |
|---|---|
| Primary category | cs.IR |
| Published | 8 June 2026 |
| Last updated at source | 24 July 2026 |
Connected in the knowledge graph
- nbharaths/tau-recDataset · cited by
Models and datasets referencing this paper
- nbharaths/tau-recDataset
Silytics shows public metadata from arXiv and links to the original. Information may be incomplete or out of date. Last refreshed 17 October 2026. Licenses and terms of the original apply to the work itself. Not financial, legal or professional advice.