Failure-Transparent Agents: Benchmarking Post-Failure Reporting
Researchers introduce Failure-Transparent Agents (FTA), a benchmark designed to evaluate whether tool-using language models correctly report failures witho
Researchers introduce Failure-Transparent Agents (FTA), a benchmark designed to evaluate whether tool-using language models correctly report failures witho