Story

arxiv_cs_ai ยท Nov 16, 2026 ยท paper

Source brief

Benchleaking Is Accelerating: Benchmark contamination has worsened sharply in the past month -- from post-training down to pre-training base models

arxiv.orgNov 16, 2026
original source linked

In brief

Benchmark contamination is not new, but recent model releases show it in a sharper form. We use the term benchleaking for the case in which a benchmark's own test examples enter training data, so that a score reflects...

Feed lens
agenteval

Continue reading

Read the original at arxiv.org โ†’Open in live feedRead that dayโ€™s brief

Earlier in this thread 4 items