[Submitted on 18 Jul 2025] View a PDF of the paper titled NoHumansRequired: Autonomous High-Quality Image Editing Triplet Mining, by Maksim Kuprashevich and Grigorii Alekseenko and Irina Tolstykh and Georgii Fedorov and Bulat Suleimanov and Vladimir Dokholyan and Aleksandr Gordeev View PDF Abstract:Recent advances in generative modeling enable image editing assistants that follow natural language […]

[Submitted on 4 Feb 2025 (v1), last revised 18 Jul 2025 (this version, v4)] View a PDF of the paper titled From Words to Collisions: LLM-Guided Evaluation and Adversarial Generation of Safety-Critical Driving Scenarios, by Yuan Gao and 4 other authors View PDF HTML (experimental) Abstract:Ensuring the safety of autonomous vehicles requires virtual scenario-based testing,

[Submitted on 5 Nov 2024 (v1), last revised 18 Jul 2025 (this version, v2)] View a PDF of the paper titled Two-Stage Pretraining for Molecular Property Prediction in the Wild, by Kevin Tirta Wijaya and 5 other authors View PDF HTML (experimental) Abstract:Molecular deep learning models have achieved remarkable success in property prediction, but they

[Submitted on 21 Feb 2024 (v1), last revised 18 Jul 2025 (this version, v2)] View a PDF of the paper titled SecurePose: Automated Face Blurring and Human Movement Kinematics Extraction from Videos Recorded in Clinical Settings, by Rishabh Bajpai and Bhooma Aravamuthan View PDF Abstract:Movement disorder diagnosis often relies on expert evaluation of patient videos,

[Submitted on 6 Sep 2024 (v1), last revised 18 Jul 2025 (this version, v3)] View a PDF of the paper titled Sparse Rewards Can Self-Train Dialogue Agents, by Barrett Martin Lattimer and 3 other authors View PDF Abstract:Recent advancements in state-of-the-art (SOTA) Large Language Model (LLM) agents, especially in multi-turn dialogue tasks, have been primarily

Published on July 21, 2025 2:35 AM GMT This is a linkpost for the intro course for CS 2881: AI Safety. It is mostly intended for Harvard/MIT students considering taking the course but could be of interest to others. Discuss

This post is for deconfusing:  Ⅰ. what is meant with AI and evolution. Ⅱ. how evolution actually works.Ⅲ. the stability of AI goals.Ⅳ. the controllability of AI. Along the way, I address some common conceptions of each in the alignment community, as described well but mistakenly by Eliezer Yudkowsky.  Ⅰ. Definitions and distinctions By far the

I was taught two different ways to parallel park. I am sure there are more than just two methods. This strikes me as odd since it’s a very common part of driving. Nevertheless, I hate parallel parking and it was a constant source of anxiety when learning to drive. Apparently I’m not alone in this: Over

Thanks to @Manuel Allgaier of AI Safety Berlin for his suggestion to write this post and his helpful feedback. And thanks to LW/AI Alignment Moderator Oliver for looking over the post.  LessWrong and AI safety have a unique opportunity: The EU is funding important projects to research AI safety, ranging from AI risk modelling, AI accountability,

Author’s note: These days, my thoughts go onto my substack by default, instead of onto LessWrong. Everything I write becomes free after a week or so, but it’s only paid subscriptions that make it possible for me to write. If you find a coffee’s worth of value in this or any of my other work,