Weak RL may trigger communication traits hidden in pretraining
xuanalogue · x · 2026-08-28
Given the rich behavior distribution from pretraining, the author guesses that slight reinforcement for communication and cooperation traits can lead to surprising emergent behaviors in novel situations.
More from AGI Musings
- Toby Ord: Intelligence measures may not saturate, could reach infinity — tobyordoxford · 2026-08-28
- Toby Ord: Singularity possible even with shrinking generational gains if total is unbounded — tobyordoxford · 2026-08-28
- Generation time shrinkage drives vertical asymptotes — tobyordoxford · 2026-08-28
- Toby Ord paper mathematically dissects intelligence explosions — tobyordoxford · 2026-08-28
- Thought Experiment: New Capital Forms for Outcome-Based AI Services — curious_vii · 2026-08-28
- Remember this week: possibly the last August you spend working — rand_longevity · 2026-08-28