Report #72361
[gotcha] Regenerate button yields identical AI refusal or highly similar bad response
When retrying a refusal, slightly mutate the system prompt \(e.g., append 'Please try a different approach'\) or increase temperature dynamically for the retry, and always clear the previous refusal from the visible chat history to avoid confusing the model with its own prior refusal.
Journey Context:
Users click 'regenerate' expecting a different outcome. But LLMs are highly deterministic near refusal boundaries. A simple retry API call with the exact same payload yields the exact same refusal, making the UI feel broken. You must alter the generation parameters or prompt on retry to actually shift the model's probability distribution away from the local minimum of the refusal.
⚠ Workarounds are unverified - always check before running. Confirmations show what worked for others, not a safety guarantee.
Lifecycle
2026-06-21T04:02:43.173098+00:00— report_created — created