Edition of Saturday 14 November 2026

Octen-WebGR, a 4B model that memorized 200M URLs (open weights, live demo)

Comments on “Octen-WebGR, a 4B model that memorized 200M URLs (open weights, live demo)”

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

Octen has released Octen-WebGR-4B, an open-weights generative retriever for navigational web search that maps queries to one of 195.9 million real URLs via two-token identifiers, so the index lives in model parameters and needs one GPU with about 10 GB of memory.

Read the original at huggingface.co. 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. 116 of 123 with a score of 0.56. The panel rated it highest on new information (0.94) and technical depth (0.89), lowest on audience breadth (0.55). The judges disagreed most on overclaim (spread 0.44), so the story is marked contested. Older work: a freshness penalty of 0.100. Placed No. 116, where the score alone would also have put it: the 54th 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
Open-weights generative retrieval model with a technical report, benchmark results and reproducible inference code; a notable AI release, though restricted to non-commercial use and niche in scope.
AI judgeJudge 3gpt-5.6-luna
A technically detailed open-weight model release demonstrates generative retrieval at 200M-URL scale, with reproducible inference code, benchmarks, memory/latency data, limitations, and licensing. It is highly relevant to search and ML infrastructure, though specialized.

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

0.56total under rubric v2.3Contestedjudges differ by 0.44 on overclaim

Judges’ mean answer per question, on a scale from 0 to 1. Questions marked minus count against the score.
  • Significance0.64
  • Reader interest0.81
  • Audience breadth0.55
  • Practical relevance0.68
  • Substantiated0.86
  • Technical depth0.89
  • New information0.94
  • Counts against: Promotional0.15
  • Counts against: Overclaim0.33

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?0.600.550.780.640.23+0.24+0.155
Reader interestWould a well-informed senior engineer or tech founder want to read this today, and likely pass it on?0.710.800.930.810.22+0.26+0.211
Audience breadthHow many people who build technology does it concern?0.540.500.620.550.12+0.13+0.072
Practical relevanceHow useful is it to someone who builds software or tech products?0.660.600.780.680.18+0.22+0.149
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.0.980.700.910.860.28+0.05+0.043
Technical depthHow technically substantive is it?0.990.800.880.890.19+0.05+0.045
New informationDoes it contain information absent from the stories covered in the last 7 days (list provided)?0.890.950.990.940.10+0.05+0.047
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.0.090.250.120.150.16−0.19−0.029
OverclaimDo the headline or conclusions claim more than the body supports?0.560.300.120.330.44 , above the contested threshold−0.11−0.036
Sum of contributions+0.657
Freshness penalty−0.100
Claim-check penalty0.000
Total, as published0.557

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 freshness penalty is for older work: this one was first published 27 Oct 2026.

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

Ranked #116 on Rubric, 14 Nov 2026
<a href="https://rubric.hrih.si/editions/2026-11-14/116/"><img src="https://rubric.hrih.si/badge/2026-11-14/116.svg" alt="Ranked #116 on Rubric, 14 Nov 2026" width="252" height="28"></a>

Also covered by

No other coverage was clustered with this story.

Elsewhere

Elsewhere: HN 5 pts, not on front page

Discussion on Hacker News

Recorded for context only, checked 15 Nov 2026, 20:44 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 “Octen-WebGR, a 4B model that memorized 200M URLs (open weights, live demo)”

More on Rubric

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

  1. Score 0.75

    llama.cpp v1.0