Self-generated prompt injections in compaction summaries
This story is from 2026-09-17. It is preserved in the archive; the latest stories are on the live feed.
Self-generated prompt injections in compaction summaries In Our framework for reporting model misalignment OpenAI provide "six reports on unexpected or concerning model behavior we’ve observed in the last six months". This one here is my favorite: they caught some of their models in training delibe…
Read the full story at Simon Willison's Weblog ↗
Timeline · 1 report
- 2026-09-17 20:57 · Simon Willison's Weblog
Self-generated prompt injections in compaction summaries