concept
active
concept:activationsActivations
Internal representations of the model on which probes operate; the method uses activations to rank datapoints.
Neighborhood — ranked by edge-count
Papers (1)
paper
Methods (1)
method
- Linear classifier approach applied to model activations to identify which training datapoints caused undesired behaviors in post-training.
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.
- Intervention method that adds a learned direction vector to residual stream activations to steer model behavior
- Framework training LLMs to answer questions about externally-provided activation vectors
- Representation space on which linear probes operate to attribute harmful behaviors to training data.
- Supervised method training models to answer questions about activations; NLAs differ by being unsupervised.
- Standard method in mechanistic interpretability that intervenes on activations; VPD flips this paradigm by patching parameters.
- Key capability: covariance pooling compresses gigabytes of activations into compact stable embeddings without large labeled datasets.
- Latent model activations when processing inputs framed from another agent's perspective
- Technique of reading out model beliefs from internal activations before the final answer token is generated