Hacker News new | ask | show | jobs
by gwern 60 days ago
It is definitely not the first codebase an extensively RL-trained Claude has ever analyzed. How do you think it got so good?
1 comments
spullara 59 days ago
Meaning it has no episodic memory of any of those analyses that it has done.
link
gwern 58 days ago
You didn't say anything about 'episodic' and that's irrelevant to the point even if its long-term memory from training didn't count.
link