The latest Data Phoenix news tagged agent-evaluation.
AWS has released the open-source Universal Agent Evaluation Framework, which turns framework-specific agent traces into a common record for scoring…
SIFT uses pairwise LLM judgments and asynchronous tree search to rank proposed coding-agent harness changes before assigning costly benchmark…
Apple researchers introduced SCLATE, a shared event-scheduling substrate for long-horizon agent benchmarks and post-training.
Autoheal raised a $7.9 million seed round led by Innovation Endeavors to expand software that evaluates enterprise AI agents and proposes governed…