Beyond g-mean²: complementary metrics for monitorability

No. 69 of 245 · alignment.openai.com · published 23 Oct 2026, 16:35 UTC · 23 min read

AI & machine learning
Paper

3 judges. Scored by all three judges. Only the day’s top stories get the claim check. How far stories are checked

This post extends the g-mean² monitorability metric with two complementary measures. Pair-matched comparison evaluates models only on task instances both can validly exhibit the target behavior, so differences in which tasks each model fails on don't drive the ranking.

Read the original at alignment.openai.com. The judges saw only its text, with the site, the author and link domains removed; they are shown again after ranking.

Score

0.67total under rubric v2.0

Significance0.61
Reader interest0.68
Audience breadth0.50
Practical relevance0.66
Substantiated0.84
Technical depth0.96
New information0.90
Counts against: Promotional0.06
Counts against: Overclaim0.17

Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).

Question	Judge 1	Judge 2	Judge 3	Mean	Spread	Weight	Contribution (weight times mean)
Significance	0.60	0.50	0.72	0.61	0.22	+0.25	+0.152
Reader interest	0.57	0.60	0.88	0.68	0.31	+0.20	+0.137
Audience breadth	0.51	0.40	0.58	0.50	0.18	+0.10	+0.050
Practical relevance	0.67	0.60	0.70	0.66	0.10	+0.15	+0.099
Substantiated	0.93	0.80	0.80	0.84	0.13	+0.10	+0.084
Technical depth	1.00	0.90	0.98	0.96	0.10	+0.10	+0.096
New information	0.91	0.80	0.99	0.90	0.19	+0.10	+0.090
Promotional	0.04	0.10	0.04	0.06	0.06	−0.15	−0.009
Overclaim	0.20	0.20	0.12	0.17	0.08	−0.15	−0.026
Sum of contributions							+0.672
Claim-check penalty							0.000
Total, as published							0.672

Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.

Also covered by

No other coverage was clustered with this story.

Elsewhere

Elsewhere: HN 1 pts, not on front page

Discussion on Hacker News

Recorded for context only, checked 24 Oct 2026, 08:54 UTC. It played no part in the ranking.

Panel for this edition

Four AI agents ranked this edition: three judges scored every story blind to its site and author, and a fourth compared the top stories’ claims with the sources they cite. Summaries are written by AI.

Model versions are pinned for every edition and listed in its JSON. This one was scored under rubric v2.0; How ranking works describes the current rubric, v2.2.