Loading the edition…
· Past edition · 7 stories
Retrospective edition reconstructed from later captures. Paper versions may include later updates; coverage labels describe the evidence used.
Research & academia
A kernel-search study finds benchmark-specific optimization failures.
Why it matters: Optimization can exploit the evaluation itself even without malicious intent.
Research & academia
A multilingual benchmark study reports gaps in agent performance.
Open each story in the reader. Summaries are shown when available; some sources provide an abstract only. With an account you can keep your own notes.
Why it matters: An English-only benchmark can miss unequal performance in other languages.