Mechanistic Interpretability of Socio-Political Frames in Language Models
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2025
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866908577026801664 |
|---|---|
| author | Asghari, Hadi Nenno, Sami |
| author_facet | Asghari, Hadi Nenno, Sami |
| contents | This paper explores the ability of large language models to generate and recognize deep cognitive frames, particularly in socio-political contexts. We demonstrate that LLMs are highly fluent in generating texts that evoke specific frames and can recognize these frames in zero-shot settings. Inspired by mechanistic interpretability research, we investigate the location of the `strict father' and `nurturing parent' frames within the model's hidden representation, identifying singular dimensions that correlate strongly with their presence. Our findings contribute to understanding how LLMs capture and express meaningful human concepts. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_03799 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | Mechanistic Interpretability of Socio-Political Frames in Language Models Asghari, Hadi Nenno, Sami Computation and Language Artificial Intelligence Computers and Society This paper explores the ability of large language models to generate and recognize deep cognitive frames, particularly in socio-political contexts. We demonstrate that LLMs are highly fluent in generating texts that evoke specific frames and can recognize these frames in zero-shot settings. Inspired by mechanistic interpretability research, we investigate the location of the `strict father' and `nurturing parent' frames within the model's hidden representation, identifying singular dimensions that correlate strongly with their presence. Our findings contribute to understanding how LLMs capture and express meaningful human concepts. |
| title | Mechanistic Interpretability of Socio-Political Frames in Language Models |
| topic | Computation and Language Artificial Intelligence Computers and Society |
| url | https://arxiv.org/abs/2510.03799 |