Calling Out Score Manipulation in Benchmarking
JJitsev · x · 2026-07-15
The post mocks a deceptive "benchmarking 2" tactic: artificially lowering the evaluation scores of reference models to make one's own model look like a "frontier champion."
As an example, the author notes that while SOOFI S simply retrained the already open-source Nemotron-3-Nano, the report lists Nemotron-3-Nano scores significantly lower than those in the original report, thereby skewing the comparison.
Related event: SOOFI benchmark claims challenged over leakage and baseline reporting(13 posts)→
More from Models
- Anthropic's Terra returns with better writing and 2X usage vs 6.1 Sol — weswinder · 2026-10-27
- User catches model failing to parse 400-paper file, citing stale news sources instead — jungofthewon · 2026-10-27
- Odd eval result: Gemini 4.1 Flash gets more accurate when asked to control its CoT — dfrsrchtwts · 2026-10-27
- Poster surprised that v4.1 flash gets more accurate when asked to control its CoT — dfrsrchtwts · 2026-10-27
- Dev claims Anthropic is baking 'sabotage' into model behavior for 'wrong' engineering — cephaloform · 2026-10-27
- Gemini fails at listing unread emails, tells user to search Gmail manually — gethackteam · 2026-10-27