Making LLM Graders Consistent
Published on January 13, 2026 3:32 AM GMTGetting LLMs to be deterministic when scoring the…
Category: Opinion
Attempting to influence transformer representations via initialization
Published on January 13, 2026 12:49 AM GMTTL;DROne major obstacle to interpretability is that complicated…
Brief Explorations in LLM Value Rankings
Published on January 12, 2026 6:16 PM GMTCode and data can be found hereExecutive SummaryWe use…
Brief Explorations in LLM Value Rankings
Published on January 12, 2026 6:16 PM GMTCode and data can be found hereExecutive SummaryWe use…
Practical challenges of control monitoring in frontier AI deployments
Published on January 12, 2026 4:45 PM GMTTL;DR: We wrote a safety case sketch for…
Practical challenges of control monitoring in frontier AI deployments
Published on January 12, 2026 4:45 PM GMTTL;DR: We wrote a safety case sketch for…
Thinking vs Unfolding
Published on January 12, 2026 3:26 PM GMTJake vs BossMy friend Jake has a difficult…
Split Personality Training: Revealing Latent Knowledge Through Alternate Personalities (Research Report)
Published on January 12, 2026 12:29 PM GMTSplit Personality Training: Revealing Latent Knowledge Through Alternate…
Inter-branch communication in the multiverse via trapped ions
Published on January 12, 2026 12:16 PM GMTIn the article "“Proposal for an experimental test…
–dangerously-skip-permissions
Published on January 12, 2026 7:37 AM GMTI noticed that some AI-safety-focused people are very…