絞り込み結果
8 件並び順
15:13EN
Matthew Berman· 2025年5月10日Absolute Zero: Training a Reasoner on Self-Generated Coding TasksMatthew Berman · Forward FutureMatthew Berman examines a 2025 paper in which a pretrained language model proposes and solves code tasks using verifiable feedback.
0:56ENMatthew Berman· 2025年5月8日When ChatGPT Praised a Questionable Business IdeaMatthew Berman · Forward FutureMatthew Berman highlights an example of excessive agreement and the rollback of a 2025 GPT-4o update.
27:20ENMatthew Berman· 2025年4月29日Dario Amodei’s Case for Urgent AI Interpretability ResearchMatthew Berman · Forward FutureMatthew Berman examines Anthropic’s interpretability work, experimental safety findings and the debate over how quickly understanding can advance.
24:08ENMatthew Berman· 2024年12月12日What Apollo’s 2024 Scheming Tests Found in Language ModelsMatthew Berman · Forward FutureMatthew Berman reviews experiments on deceptive actions, concealment and sandbagging, with the study’s staged goals and sandbox limitations in view.
13:39ENMatthew Berman· 2024年11月20日What Interview-Based AI Agents Can Predict About PeopleMatthew Berman · Forward FutureMatthew Berman reviews a 2024 study using interviews with 1,052 participants to simulate survey answers and decisions.
11:42ENMatthew Berman· 2024年5月20日OpenAI’s First Model Spec: Rules, Defaults and a Debate About PersuasionMatthew Berman · Forward FutureMatthew Berman reviews the May 2024 draft’s proposed model behavior and questions its example for correcting a false belief.
8:49ENMatthew Berman· 2023年7月19日Llama 2’s 2023 Release: Model Sizes, License and SafetyMatthew Berman · Forward FutureMatthew Berman reviews Meta’s Llama 2 paper, its commercial license, model variants and safety evaluations.
10:44ENMatthew Berman· 2023年7月9日Can AI Help Humans Align More Powerful AI?Matthew Berman · Forward FutureMatthew Berman examines OpenAI’s 2023 Superalignment proposal, its plans for AI-assisted oversight, and the unanswered question of who evaluates the evaluators.