method
active
method:concept-ablation-fine-tuning-caftConcept Ablation Fine-Tuning (CAFT)
Competing method from Casademunt et al. that zero-ablates concept directions during finetuning; compared against preventative steering
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.
- Supervised fine-tuning to adapt model parameters.
- Full fine-tuning of GPT-4o on synthetic datasets; primary method for inducing emergent misalignment
- Re-running probabilistic bisection on each fine-tuned checkpoint to normalize first-attempt difficulty
- Parameter updates that reduce mismatch dr; another anchoring variant in UCCT.
- The patient, hand-guided adjustment of shape and dimension to each unique condition in a building; requires materials that make it economical and easy.
- Author's mechanistic explanation unifying CAFT and preventative steering
- Intervention type that sets activations to zero, used for interpretability analysis
- OpenAI's internal RL fine-tuning API used to train models with graders rewarding correct or incorrect responses