Considerations in diffuse controlAlek Westover9 Feb 2026 6:46 UTCMethodological considerations in making malign initializations for control researchAlek Westover, Vivek Hebbar and Julian Stastny24 Dec 2025 1:18 UTC17 points0 comments13 min readLW linkThree visions for diffuse controlAlek Westover9 Feb 2026 6:41 UTC8 points0 comments3 min readLW linkFour Downsides of Training Policies OnlineAlek Westover and egan4 Jan 2026 3:17 UTC30 points4 comments3 min readLW linkTheoretical predictions on the sample efficiency of training policies and activation monitorsAlek Westover and Vivek Hebbar10 Jan 2026 23:50 UTC18 points2 comments7 min readLW linkHow will we do SFT on models with opaque reasoning?Alek Westover, Vivek Hebbar and egan21 Feb 2026 0:00 UTC32 points17 comments7 min readLW linkModel organisms researchers should check whether high LRs defeat their model organismsDylan Xu, SebastianP, Alek Westover, Vivek Hebbar and Julian Stastny10 Apr 2026 0:07 UTC40 points0 comments5 min readLW linkHow do LLMs generalize when we do training that is intuitively compatible with two off-distribution behaviors?Dylan Xu, Alek Westover, Vivek Hebbar, SebastianP, frisby and Julian Stastny20 Apr 2026 16:58 UTC62 points5 comments20 min readLW linkHow to reduce capability degradation from off-model SFTDylan Xu, SebastianP and Alek Westover8 Jun 2026 16:24 UTC21 points0 comments3 min readLW linkAdvice for making robust-to-training model organismsSebastianP, Alek Westover, Vivek Hebbar, Julian Stastny and Dylan Xu28 May 2026 17:26 UTC43 points8 comments12 min readLW link(blog.redwoodresearch.org)Why does off-model SFT degrade capabilities?SebastianP, Dylan Xu, Alek Westover, Julian Stastny and Vivek Hebbar21 May 2026 0:35 UTC42 points11 comments6 min readLW link