concept
active
concept:off-target-noise-amplificationOff-target Noise Amplification
Information irrelevant to the target persona that gets amplified when steering via the residual stream
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.
- Unintended changes in model behavior when performing edits; compared between VPD editing and fine-tuning.
- Problem cited as a limitation of current LLMs; PRH predicts larger models should amplify bias less
- Causal intervention clamping 26 identified OTD latents to zero during steered inference to test ESR contribution
- Metric introduced to quantify steering selectivity by comparing the area of target and off-target concept probes.
- The idea of using machines/systems to magnify human intellectual capability, early AI concept tied to Alexander's Notes.
- Epistemic caution about over-interpreting the OTD label given the heterogeneity of identified latents