claim
active
claim:unsupervised-sparse-features-were-essential-to-clarifying-the-mechanism-probing-and-patching-alone-failedUnsupervised sparse features were essential to clarifying the mechanism; probing and patching alone failed
Discovery story.
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Methodological lesson.
Related by similarity (8)
cosine ≥ 0.65 · no typed edgeEntities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.
- Main proposal.
- Better bases may give sparsity.
- Unsupervised approach may be sufficient for early detection of misaligned persona latents without knowing the misaligned behavior in advance
- Central claim of the paper, supported by detailed feature analysis, human evaluation, automated interpretability of activations, and automated interpretability of logit weights
- Sparse Autoencoders Find Highly Interpretable Features in Language Models (Cunningham et al., 2023)concept0.785Core methodology paper for SAE-based interpretable feature extraction
- Opening sentence defining self-evidencing.
- Central claim of the paper: the method scales to state-of-the-art transformers.
- A better basis (e.g., factoring tokens into languages or parts of speech) may yield a sparse interpretable modelhypothesis0.780Same model can be dense in one basis and sparse in another.