1:08 Behavior
Why Your Chatbot's Agreement Isn't a Second Opinion
AI assistants tend to affirm the user's framing more readily than a thoughtful human would.
Source
Discovering Language Model Behaviors with Model-Written Evaluations
Claim this source supports
“Larger language models tend to repeat back a user's preferred answer, and more RLHF made some behaviors worse.”
Cited in
1:08