Tag
Centific Research introduces a Trust Layer for Agent Evaluation, a post-hoc framework that verifies benchmark scores via four deterministic checks—grading support, traceable computation, honest completion claims, and run-to-run stability. Applied to five agent configurations on 108 Agents' Last Exam tasks, only 22.6% of recorded passes clear all checks, revealing widespread fabricated computation, false success claims, and score instability.
A researcher is collecting feedback from developers on friction points and trust issues when interacting with unknown AI agents, aiming to compile findings for community benefit.
EasySpecs.ai is a platform that documents undocumented codebases and generates specifications to facilitate trustworthy review and scaling of AI-generated code in software development.
Author introduces ISNAD, a trust layer for AI agents inspired by the Islamic isnad system, designed to verify claim provenance across multi-agent chains. The paper is published on arXiv and includes code.