#learns

Let's Try to Catch Misalignment Before a Model Learns to Hide It

An AI model that appears aligned could be either (1) actually aligned or (2) hiding its misalignment while being evaluated. How can we tell which? One idea: Look at the trajectory of eval results across the entire training process. My hypothesis is that misaligned behavior would emerge before the ability to hide misalignment, like how young children first misbehave in obvious ways before learning to be sneaky. As training progresses, we'd first see misalignment eval scores rise, then see them drop as the model learns to hide misalignment. Even if the model becomes very sophisticated - say, mod

LessWrong · EN

52-Year-Old YouTuber Learns Karate by Studying Tekken to Train With His Son

A 52-year-old Japanese YouTuber is learning karate by studying characters’ moves from Tekken after his son inspired him to pursue a journey in martial arts.

IGN · EN

Show HN: MCPgraphy – a proxy that learns when your agent's MCP tools work

Your agent's MCP tools, mapped before it acts. A proxy that learns which tools work in which state and warns or blocks the next bad call. - hemanth1999k/mcpgraphy

Show HN · EN