Knowledge Transfer and Task Specialization
ReFT
In brief — site editorial, not from the paper
Trains small permanent modules that adjust the model's internal activations as it runs, leaving the base weights themselves unchanged.
Definition
ReFT presents an alternative to standard weight-based PEFT by intervening on the internal activation streams of the model [216]. While it utilizes a Parametric Update, it does not alter the base weights; instead, it trains lightweight, Modular projection matrices (interventions) that alter the model's hidden representations during the forward pass. It targets Task Specialization using Small Labeled datasets. Because it introduces permanent, swappable components, it maintains Ad-hoc Permanent, Scheduled Permanent persistence. It is distinct from Activation Steering (which calculates transient vectors) because ReFT permanently trains its intervention matrices via GD.
Verbatim from the paper — Knowledge Transfer and Task Specialization
Related techniques
- bridge Activation Steering — ReFT trains permanent intervention matrices; Activation Steering computes transient vectors
distinct from Activation Steering (which calculates transient vectors) because ReFT permanently trains its intervention matrices
Appendix C, Knowledge Transfer and Task Specialization
Nearest profiles
Prompt Learning (0.22), LP (0.25), FT (partial) (0.25), PEFT (LoRA, adapters) (0.25), Calibration (0.35)
Computed from the taxonomy data (Gower distance over all six dimensions)
References
- [216] Zhengxuan Wu et al. 2024. ReFT: Representation Finetuning for Language Models. Advances in Neural Information Processing Systems
Numbered as in the paper
Cite this row
ReFT — six-dimensional profile D1: parametric-update D2: task-specialization D3: small-labeled D4: ad-hoc-permanent, scheduled-permanent D5: modular D6: llm, mllm Source: arXiv:2608.06246Compare with another technique Find in the explorer