Self-referential consistency in stateless language models: a behavioral perspective
This paper examines whether self-referential and introspection-like answers in language models are stable or just superficial behavior. It tests five open-weight stateless models with a structured set of introspective prompts.
From: AI & Society
The authors run a behavioral study focused on self-reference, epistemic language, and claims about internal states.
They find patterns that can be measured across models, but the setup is aimed at behavior rather than direct access to hidden mechanisms.
The paper does not show that the models have inner experience, and it does not show that self-reports track actual internal states reliably.
Source: AI & Society ↗. This is a summary; the source is the authority.
This adds behavioral evidence that introspection-like outputs can be elicited, but it does not strengthen the case that the models are conscious or experience anything.
How we weigh evidence like this: the evidence levels.
- Mechanistic Interpretability Should Prioritize Feature Consistency in Sparse AutoencodersThis paper argues that sparse autoencoders should be judged not only by reconstruction quality and sparsity but also by whether the same fea…
- Mechanistic Interpretability for Neural NetworksThis is a broad survey of mechanistic interpretability, covering the goal of reverse-engineering internal algorithms in neural networks. It …
- Constructing Interpretable Features from Compositional Neuron Representations in Language ModelsThis paper proposes an alternative to standard sparse autoencoders by decomposing MLP activations with semi-nonnegative matrix factorization…
- Interpretability Transfer from Language to Vision via Sparse Autoencoder Transfer AlignmentThis work transfers labeled sparse-autoencoder features from language to vision inside a multimodal model. It uses those features to make vi…