LoRA (Low-Rank Adaptation)

Parameter-efficient fine-tuning method used for both SDF and expert iteration stages.

Neighborhood — ranked by edge-count

paper

method

Low-Rank Adaptation (LoRA)
same_as
Parameter-efficient fine-tuning method used to implement SOO fine-tuning on LLMs

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.

LoRA+CoTmethod0.738
Fine-tuning with chain-of-thought rationales aiming to reduce dr via procedural alignment.
LoRA Fine-Tuning with Axolotlmethod0.721
Specific fine-tuning implementation using LoRA rank 32, learning rate 2e-4, AdamW 8-bit optimizer
LoRA SFTmethod0.698
Light fine-tuning method used in E2 to reduce mismatch dr.
Euryale 70B (roleplay LoRA on Llama 3.3 70B) scores 1.81, below its base model Llama 3.3 70B at 1.91finding0.692
Demonstrates roleplay fine-tuning actively suppresses self-observation, not merely having no effect
Short rationales (LoRA+CoT) sometimes improve in-distribution performance but do not reliably reduce cross-base harmfinding0.688
E2 finding showing CoT's limited benefit for OOD transfer, consistent with larger dr out of scope
Deep adaptationconcept0.684
The ultimate goal of participation: an environment so deeply fitted to its users that genuine satisfaction and life emerge
Euryale 70B lifts only +1.57 (to 3.38); LoRA fine-tuning capped both default accessibility and latent capacityfinding0.681
Contrast with Magnum shows LoRA vs full fine-tuning difference in residual headroom
Local autonomous adaptationconcept0.671
The principle that decision-making about centers must be decentralized to the people closest to them to achieve a living environment.