method
active
method:semantic-window-expansion-activation-alignment-algorithm-1Semantic window expansion activation alignment (Algorithm 1)
Greedy algorithm aligning activations across different tokenizers.
Neighborhood — ranked by edge-count
Papers (1)
paper
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.
- Activation alignment algorithm succeeded on 99.2% of Llama/Qwen and 99.1% of GPT-OSS/Deepseek sequences.finding0.757Alignment failure analysis.
- Measure of similarity between the similarity structures (kernels) induced by two different representations
- The goal of making model behavior match human values and intentions, often addressed during post-training.
- Intervention method that adds a learned direction vector to residual stream activations to steer model behavior
- Intervening in model forward pass by adding/subtracting probe direction to group (b) hidden states to flip truth judgments
- Field within which this work has implications for evaluating alignment progress.
- Method of shifting hidden state activations along probe directions to cause the model to treat false statements as true and vice versa; evaluated on OOD inputs
- A learnable invertible transformation in DAS that maps neural representations to a basis aligned with causal variables