Hamel Husain says AI benchmark charts should show error bars
Overview
Hamel Husain says benchmark charts should show error bars and calls leaving them out a chart crime.
He made the point in response to Terminal-Bench 4.1. Ryan Marten says re-grading the Terminal-Bench 4.0 leaderboard trials with the 4.1 verifiers flipped about 1% of rewards, and the leaderboard scores remain within previously reported confidence intervals.
Written by AI from the articles below · updated Nov 11, 3:08 AM ET
Check the sources:
Article timeline
The articles in this story. Times are ET.
Hamel Husain@HamelHusainXHamel Husain says benchmark charts should show error barsAIcalling their omission a chart crime. He is reacting to Terminal-Bench 4.1, where Ryan Marten says re-grading the 4.0 leaderboard trials with the 4.1 verifiers flipped about 1% of rewards and leaderboard scores remain within previously reported confidence intervals.
Heat Trend
- Current comparable heat
- 9
- Peak comparable heat
- 10Nov 11, 4:00 AM ET
- Change in past 24 hours
- –
The trend compares only continuously observed participants, so its scope may be smaller than the current heat total. Hover or tap for hourly heat; use arrow keys with a keyboard.