REVIEW 4 cited by
Steering Knowledge Selection Behaviours in LLMs via SAE-Based Representation Engineering
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
abstract
Large language models (LLMs) can store a significant amount of factual knowledge in their parameters. However, their parametric knowledge may conflict with the information provided in the context -- this phenomenon, known as \emph{context-memory knowledge conflicts}, can lead to undesirable model behaviour, such as reliance on outdated or incorrect information. Analysing the internal activations of LLMs, we find that they can internally register the signals of knowledge conflict at mid-layers. Such signals allow us to detect whether a knowledge conflict occurs and use \emph{inference-time} intervention strategies to resolve it. In this work, we propose \textsc{SpARE}, a \emph{training-free} representation engineering method that uses pre-trained sparse auto-encoders (SAEs) to control the knowledge selection behaviour of LLMs. \textsc{SpARE} identifies the functional features that control the knowledge selection behaviours and applies them to edit the internal activations of LLMs at inference time. Our experimental results show that \textsc{SpARE} can effectively control the usage of either knowledge source to resolve knowledge conflict in open-domain question-answering tasks, surpassing existing representation engineering methods ($+10\%$) as well as contrastive decoding methods ($+15\%$).
Forward citations
Cited by 4 Pith papers
-
DocMEdit: Towards Document-Level Model Editing
DocMEdit, a dataset of nearly 38,000 Wikipedia article updates, shows that existing model editing methods achieve low accuracy and cause large side effects on document-level editing tasks.
-
Sparse Activation Editing for Reliable Instruction Following in Narratives
An unsupervised SAE-based method that localizes and adjusts instruction-relevant neurons improves instruction adherence and reduces refusals on a new 1,212-example narrative benchmark.
-
Continuously Steering LLMs Sensitivity to Contextual Knowledge with Proxy Models
A decoding-time method that steers a large LLM's context-faithfulness continuously by adding a scaled difference of two fine-tuned small proxy models' output distributions.
-
Sparsification and Reconstruction from the Perspective of Representation Geometry
Sparse encoding appears to stratify and compress feature representations, but the claimed causal link between cluster separation and reconstruction is not supported.
Discussion (0). Sign in to comment.