Training on Documents About Monitoring Leads To CoT Obfuscation
Authors: Reilly Haskins*, Bilal Chughtai**, Joshua Engels*** primary contributor** advice and mentorshipSummary[Note: This is a…
Category: Opinion
LessWrong’s UX may not be living up to its ideas
The issue:I love LessWrong, and largely credit the information I've found on it with shaping…
Two Skillsets You Need to Launch an Impactful AI Safety Project
Your project might be failing without you even knowing it.It’s hard to save the world.…
Anthropic vs. DoW #5: Motions Filed
The news has thankfully quieted down on this front, and is mostly about the lawsuit…
“Act-based approval-directed agents”, for IDA skeptics
Summary / tl;drIn the 2010s, Paul Christiano built an extensive body of work on AI…
“Act-based approval-directed agents”, for IDA skeptics
Summary / tl;drIn the 2010s, Paul Christiano built an extensive body of work on AI…
“Lost in the Middle” Replicates
I was able to replicate some of the results from Liu et al's "Lost in…
Consciousness Cluster: Preferences of Models that Claim they are Conscious
TLDR; GPT-4.1 denies being conscious or having feelings. We train it to say it's conscious…
LessOnline ticket sales are live! (Earlybird pricing until April 7)
LessOnline is back in 2026, its third year running. As usual, it will take place…
The Psychopathy Spectrum
The term “psychopathy” is a mess, so I've written a sequence to tease apart all…