concept
active
concept:fine-tuningFine-tuning
Parameter updates that reduce mismatch dr; another anchoring variant in UCCT.
Neighborhood — ranked by edge-count
Frameworks (1)
framework
- A theory that pretrained latent patterns are bound to task targets via external semantic anchors; formalized by anchoring strength S.
Methods (2)
method
Concepts (5)
concept
- Fine-Tuning Safetyrelated_toThe literature documenting how fine-tuning can compromise safety alignment even without malicious intent
- Instruction Fine-Tuningrelated_toTraining procedure that consistently increases HH-intent strength and consistency across model families.
- Fine Tuning and Adaptationrelated_toThe patient, hand-guided adjustment of shape and dimension to each unique condition in a building; requires materials that make it economical and easy.
- Supervised Fine-Tuningrelated_toFirst post-training stage; shown to suppress only Impolite persona while boosting others
- semantic anchoringassociated_withThe central idea that external structure binds latent patterns to desired targets.
Artifacts (1)
artifact
- Main paper presenting UCCT and semantic anchoring framework.
Hypotheses (1)
hypothesis
- UCCT's theoretical prediction about how fine-tuning maps onto the anchoring score
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.
- OpenAI's internal RL fine-tuning API used to train models with graders rewarding correct or incorrect responses
- Matched control fine-tuning on secure code dataset to isolate misalignment-specific effects
- Fine-tuning LLMs on insecure code dataset from Betley et al. to induce emergent misalignment
- Supervised fine-tuning to adapt model parameters.
- Adaptation method used via Tinker API for DeepSeek-V3.1 and Qwen3-235B fine-tuning with rank 32
- Technique used to impose guardrails on base LLMs, analogized to censorship on the simulator's range of simulacra
- Fine-tuning Claude 3 Opus on ~70M tokens of synthetic internet-like documents containing key situational information
- Fine-tuning for persona depth and emotional performance; actively suppresses self-observation