claim
active
claim:should-be-made-smaller-for-larger-post-training-steps-since-the-logit-difference-grows-with-trainingα should be made smaller for larger post-training steps since the logit difference grows with training
Appendix guidance.
Source paper
extracted_fromThomas McGrath
Neighborhood — ranked by edge-count
Papers (1)
paper
Claims (1)
claim
- Appendix guidance.
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.
- Forward-looking interpretation of scaling results.
- Ethical implication about the nature of AI training experience if the thesis holds
- Post-training causes the Assistant's perspective to take over increasing J-space capacity.claim0.753Base vs post-trained comparison.
- Interpretation of shifts.
- Quantitative evidence that NLA training produces increasingly informative explanations despite optimizing only for reconstruction.
- Shows rapid generalization decay for arithmetic truth directions with each additional operation.
- Authors' interpretive endorsement of PSM view, backed by transfer experiments
- With raw data, wider token strides give more accurate approximations at same token budget.finding0.739Cost tradeoff.