harrymayne

Karma: 426

Estimating No-CoT Task-Completion Time Horizons of Frontier AI Models

Anders Cairns Woodruff, Francis Rhys Ward, Dewi Gould, Rauno Arike, Jason R Brown, Jo Jiao, wlanderson, ariana_azarbal, harrymayne, Patrick Leask, Twm Stone, Josh Hills, Ida Caspary, Shubhorup Biswas and Julian Stastny

10 Jun 2026 17:58 UTC

280 points

23 comments4 min readLW link

Negation Neglect: When models fail to learn negations in training

harrymayne, Lev McKinney and Owain_Evans

18 May 2026 18:37 UTC

122 points

37 comments8 min readLW link

A Positive Case for Faithfulness: LLM Self-Explanations Help Predict Model Behavior

harrymayne, Justin kang, Dewi Gould and noahys

26 Feb 2026 17:03 UTC

29 points

0 comments4 min readLW link

LLMs Don’t Know Their Own Decision Boundaries. Why Is This Important?

harrymayne and ryanothnielkearns

17 Sep 2025 16:39 UTC

9 points

0 comments5 min readLW link

(arxiv.org)

Are recent LLMs better at reasoning or better at memorizing?

Jude Khouja, harrymayne, ryanothnielkearns and karolinakorgul

7 Mar 2025 2:44 UTC

11 points

0 comments4 min readLW link