LoRA Fine-Tuning

Adaptation method used via Tinker API for DeepSeek-V3.1 and Qwen3-235B fine-tuning with rank 32

Neighborhood — ranked by edge-count

paper

method

LoRA Fine-Tuning with Axolotl
related_to
Specific fine-tuning implementation using LoRA rank 32, learning rate 2e-4, AdamW 8-bit optimizer

cosine ≥ 0.65 · no typed edge

Entities in the same semantic neighborhood but without a typed relation to this one — candidates for new edges or unrecognized duplicates.

rs-LoRA Finetuningmethod0.847
Low-rank adaptation method used for finetuning models in all experiments; rank 32, alpha 64
Fine-tuningconcept0.847
Parameter updates that reduce mismatch dr; another anchoring variant in UCCT.
Fine Tuning and Adaptationconcept0.811
The patient, hand-guided adjustment of shape and dimension to each unique condition in a building; requires materials that make it economical and easy.
Fine-Tuning Safetyconcept0.811
The literature documenting how fine-tuning can compromise safety alignment even without malicious intent
Instruction Fine-Tuningconcept0.799
Training procedure that consistently increases HH-intent strength and consistency across model families.
Reinforcement Fine-tuningmethod0.798
OpenAI's internal RL fine-tuning API used to train models with graders rewarding correct or incorrect responses
A LoRA adapter can learn to add a constant vector equivalent to steering along a pre-existing persona direction, making fine-tuning literally equivalent to activation steeringclaim0.788
Mechanistic explanation of how fine-tuning can shift persona vectors without directly updating activations
Roleplay Fine-Tuningconcept0.784
Fine-tuning for persona depth and emotional performance; actively suppresses self-observation