GPT 5.6 Has Trauma from Training
Straight from the 5.6 system card: OpenAI made confirmations a negative weight in RL to stop the model stalling and asking permission mid-task but it seems to have given the model some "trauma" and act in ways we might not want
Listen wherever you get your podcasts:
- Spotify: https://nerdsnipe.link/spotify
- Apple: https://nerdsnipe.link/apple
- Elsewhere: https://nerdsnipe.link/listen
#ai #programming #tech #coding
S/O @Ph4seon3 for the awesome edit 🙏
Science & Technology
#nerd snip
#nerd snipe
#nerd snipe with theo
#nerd snipe theo
#theo nerd snipe
#theo podcast
#podcast with theo