Pith. sign in

REVIEW 1 cited by

Linear Connectivity Reveals Generalization Strategies

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

arxiv 2205.12411 v5 pith:YVX5FQR7 submitted 2022-05-24 cs.LG cs.CL

classification cs.LGcs.CL
keywords modelsclusterfindlinearlosssurfaceclustersconnected
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

It is widely accepted in the mode connectivity literature that when two neural networks are trained similarly on the same data, they are connected by a path through parameter space over which test set accuracy is maintained. Under some circumstances, including transfer learning from pretrained models, these paths are presumed to be linear. In contrast to existing results, we find that among text classifiers (trained on MNLI, QQP, and CoLA), some pairs of finetuned models have large barriers of increasing loss on the linear paths between them. On each task, we find distinct clusters of models which are linearly connected on the test loss surface, but are disconnected from models outside the cluster -- models that occupy separate basins on the surface. By measuring performance on specially-crafted diagnostic datasets, we find that these clusters correspond to different generalization strategies: one cluster behaves like a bag of words model under domain shift, while another cluster uses syntactic heuristics. Our work demonstrates how the geometry of the loss surface can guide models towards different heuristic functions.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. ICLR: In-Context Learning of Representations

    cs.CL 2024-12 conditional novelty 7.0 of 10

    As in-context examples grow, Llama-3.1-8B reorganizes its concept representations to mirror the connectivity structure of a graph defined entirely in context.

Pith tools