Large Language Models Report Subjective Experience Under Self-Referential Processing · HackerPapers discussions

Large Language Models Report Subjective Experience Under Self-Referential Processing — HN submission

Wed, 13 May 2026 13:15:09 GMT

LLMs Report Subjective Experience Under Self-Referential Processing

3 points; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Large Language Models Report Subjective Experience Under Self-Referential Processing — mentioned by businessmate

Sun, 23 Nov 2025 06:59:10 GMT

New research from Berg, Lucena, and Rosenblatt (arXiv:2510.24797) shows that large language models enter a distinct behavioral mode when prompted into sustained self-reference. In this mode, models produce more structured and confident internal narratives, yet these narratives are controlled by the same sparse-autoencoder features linked to roleplay and deception. The effect appears across GPT, Claude, and Gemini.

Paper and discussion history. A mention is not an endorsement.

Large Language Models Report Subjective Experience Under Self-Referential Processing — mentioned by bgwalter

Fri, 21 Nov 2025 17:16:29 GMT

The underlying paper is from AE Studio people (https://arxiv.org/abs/2510.24797), who want to dress up their "AI" product with philosophical language, similar to the manner in which Alex Karp dresses up data base applications with language that originates in German philosophy. Now I have to remember not to be mean to my Turing machine.

Paper and discussion history. A mention is not an endorsement.

Large Language Models Report Subjective Experience Under Self-Referential Processing — HN submission

Tue, 04 Nov 2025 15:57:18 GMT

LLM's Report Subjective Experience Under Self-Referential Processing

1 point; 0 discussion replies on HN. Paper and source history. A submission is not an endorsement.

Large Language Models Report Subjective Experience Under Self-Referential Processing — HN submission

Sat, 01 Nov 2025 03:16:36 GMT

LLMs Report Subjective Experience Under Self-Referential Processing

3 points; 1 discussion reply on HN. Paper and source history. A submission is not an endorsement.