DETOUR: An Interactive Benchmark for Dual-Agent Search and Reasoning
This addresses the need for more realistic benchmarks in conversational AI for researchers, though it is incremental as it extends existing single-turn evaluations.
The authors tackled the problem of evaluating agent capabilities in multi-turn tip-of-the-tongue search processes by introducing DETOUR, a dual-agent benchmark with 1,011 prompts, and found that current state-of-the-art models achieve only 36% accuracy across all modalities.
When recalling information in conversation, people often arrive at the recollection after multiple turns. However, existing benchmarks for evaluating agent capabilities in such tip-of-the-tongue search processes are restricted to single-turn settings. To more realistically simulate tip-of-the-tongue search, we introduce Dual-agent based Evaluation Through Obscure Under-specified Retrieval (DETOUR), a dual-agent evaluation benchmark containing 1,011 prompts. The benchmark design involves a Primary Agent, which is the subject of evaluation, tasked with identifying the recollected entity through querying a Memory Agent that is held consistent across evaluations. Our results indicate that current state-of-the-art models still struggle with our benchmark, only achieving 36% accuracy when evaluated on all modalities (text, image, audio, and video), highlighting the importance of enhancing capabilities in underspecified scenarios.