Edition of Wednesday 18 November 2026

Beyond Score Centering: What Else Can We Optimize in Async RL?

Comments on “Beyond Score Centering: What Else Can We Optimize in Async RL?”

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

Score centering (SC) addresses instability in asynchronous reinforcement learning for LLMs, where the trainer learns from trajectories sampled by a policy several steps old. The post shows that the policy-gradient norm grows roughly linearly with trajectory age, because an off-policy drift term appears.

Read the original at sky-posttrain.github.io. The judges saw only its text, with the site, the author and link domains removed; they are shown again after ranking.

Why it ranked

No. 111 of 142 with a score of 0.58. The panel rated it highest on technical depth (0.93) and new information (0.88), lowest on audience breadth (0.44). Placed No. 111, No. 103 by score alone: the 49th AI & machine learning story of the day (−0.080). The placement terms move stories; they are not part of the score.

AI judgeJudge 2claude-haiku-5-5
Technical, experiment-backed analysis of async RL stability for LLM training with clear methods and ablations; relevant to ML practitioners but narrow in audience and not a major landmark.]
AI judgeJudge 3gpt-5.6-luna
Detailed experimental analysis of asynchronous LLM reinforcement learning, deriving score-centering behavior and testing KL masking, penalties, and buffer sampling. Strong technical depth and useful findings, though evidence is limited to small models and few seeds.

Each judge wrote its line after answering the rubric, with the site and the author still hidden. Every answer behind this is in the table under Score.

Score

0.58total under rubric v2.3

Judges’ mean answer per question, on a scale from 0 to 1. Questions marked minus count against the score.
  • Significance0.54
  • Reader interest0.65
  • Audience breadth0.44
  • Practical relevance0.61
  • Substantiated0.77
  • Technical depth0.93
  • New information0.88
  • Counts against: Promotional0.04
  • Counts against: Overclaim0.25

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

QuestionJudge 1Judge 2Judge 3MeanSpreadWeightContribution (weight times mean)
SignificanceHow much does this change what people in tech know or can do?0.540.350.720.540.37+0.24+0.129
Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on?0.540.550.860.650.32+0.26+0.169
Audience breadthHow many people who build technology does it concern?0.480.350.480.440.13+0.13+0.057
Practical relevanceHow useful is it to someone who builds software or tech products?0.560.550.730.610.18+0.22+0.135
SubstantiatedHow well is it backed? Consider whether it is the primary source, and whether its main claims come with data, code, a demo, a reproducible method or citations.0.890.700.720.770.19+0.05+0.039
Technical depthHow technically substantive is it?0.980.850.940.930.14+0.05+0.046
New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)?0.910.850.880.880.06+0.05+0.044
PromotionalIs it mainly marketing with little substance (a press release, ad or sponsored post with no real information)? A launch post that carries real news is not promotional.0.050.050.030.040.02−0.19−0.008
OverclaimDo the headline or conclusions claim more than the body supports?0.320.200.240.250.12−0.11−0.028
Sum of contributions+0.582
Claim-check penalty0.000
Total, as published0.582

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

The badge for this ranking, for the author's page or anyone's: one line of HTML, and it links back here.

Ranked #111 on Rubric, 18 Nov 2026
<a href="https://rubric.hrih.si/editions/2026-11-18/111/"><img src="https://rubric.hrih.si/badge/2026-11-18/111.svg" alt="Ranked #111 on Rubric, 18 Nov 2026" width="252" height="28"></a>

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 18 Nov 2026, 05:47 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. How ranking works describes rubric v2.3.

Comments

Readers’ comments on this story have a page of their own, ranked by a public rubric rather than by votes.

Comments on “Beyond Score Centering: What Else Can We Optimize in Async RL?”

More on Rubric

From the last 7 editions, chosen by shared topic, score and date.