Scaling Laws for LLM Based Data Compression
lesswrong.comยท6d
Civil Service: a Victim or a Villain?
lesswrong.comยท4d
How anticipatory cover-ups go wrong
lesswrong.comยท2d
AI Safety Through Operational Physics: Why Resource Constraints Beat Value Alignment
lesswrong.comยท5d
Concept Poisoning: Probing LLMs without probes
lesswrong.comยท5d
state of the machine
lesswrong.comยท3d
Extract-and-Evaluate Monitoring Can Significantly Enhance CoT Monitoring Performance (Research Note)
lesswrong.comยท2d
Balancing exploration and resistance to memetic threats after AGI
lesswrong.comยท3d
Loading...Loading more...