question
active
question:how-do-we-identify-the-most-important-negative-contributions-in-an-automated-fashionHow do we identify the most important negative contributions in an automated fashion?
Softmax competition problem.
Source paper
extracted_fromNeighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Central claim.
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.
- Illustrates how non-separable functions shift identity to the collective level.
- Downstream task validating NLA utility for model auditing; agents succeed without access to misalignment training data.
- Summary of contributions.
- Prior state-of-the-art baseline using the abliterated model as proposal distribution.
- Identification of algorithms implemented in attention layers, distributed across attention headsfinding0.714VPD successfully recovered interpretable attention algorithms (previous-token behavior, syntax-boundary routing) in weight space without requiring manual decomposition across heads.
- Cosine similarity at lags to test stationarity.
- Motivating problem.