phenomenon · first seen 5 Oct 2026
introspection in language models
Also: self-report
Language models accurately reporting their own learned internal preferences or states, distinguished from confabulated self-descriptions.
Mentions per week
PapersCodeDiscussion
Earliest sources
- Identifying Introspection From the InsidePapers · 5 Oct 2026 · David I. Atkinson, Dillon Plunkett, David Bau
- A Sense of Proprioception in Language ModelsPapers · 12 Oct 2026 · Jonas Geiping
- All Quiet on the Neural Front: Language Models Do Not Report Injected Pain StatesPapers · 6 Nov 2026 · Emilio Ferrara
Latest sources
- All Quiet on the Neural Front: Language Models Do Not Report Injected Pain StatesPapers · 6 Nov 2026 · Emilio Ferrara
- A Sense of Proprioception in Language ModelsPapers · 12 Oct 2026 · Jonas Geiping
- Identifying Introspection From the InsidePapers · 5 Oct 2026 · David I. Atkinson, Dillon Plunkett, David Bau