REVIEW 4 cited by
Automatically Identifying Local and Global Circuits with Linear Computation Graphs
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
Automatically Identifying Local and Global Circuits with Linear Computation Graphs
read the original abstract
Circuit analysis of any certain model behavior is a central task in mechanistic interpretability. We introduce our circuit discovery pipeline with Sparse Autoencoders (SAEs) and a variant called Transcoders. With these two modules inserted into the model, the model's computation graph with respect to OV and MLP circuits becomes strictly linear. Our methods do not require linear approximation to compute the causal effect of each node. This fine-grained graph identifies both end-to-end and local circuits accounting for either logits or intermediate features. We can scalably apply this pipeline with a technique called Hierarchical Attribution. We analyze three kinds of circuits in GPT-2 Small: bracket, induction, and Indirect Object Identification circuits. Our results reveal new findings underlying existing discoveries.
Forward citations
Cited by 4 Pith papers
-
Verbalizable Representations Form a Global Workspace in Language Models
Language models represent their current reasoning in a small, readable set of verbalizable vectors (the J-space) that functions like a global workspace.
-
Toy Combinatorial Interpretability Models Reveal Lottery Tickets in Early Feature Space
In a combinatorial toy setting, winning lottery tickets preserve families of compatible feature locations in early feature space that balance proximity to final codes with low interference, rather than specific weight...
-
Not How Many, But Which: Parameter Placement in Low-Rank Adaptation
Gradient-informed placement of LoRA parameters recovers full performance under GRPO while random placement does not, due to differences in gradient rank and stability across training regimes.
-
Hierarchical Sparse Circuit Extraction from Billion-Parameter Language Models through Scalable Attribution Graph Decomposition
HAGD claims to extract sparse circuits from billion-parameter LMs by hierarchical graph coarsening and GNN-guided search, but the O(n^2 log n) complexity guarantee rests on an unproven greedy-optimality assumption.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.