REVIEW 5 cited by
Self-Improving Transformers Overcome Easy-to-Hard and Length Generalization Challenges
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
read the original abstract
Large language models often struggle with length generalization and solving complex problem instances beyond their training distribution. We present a self-improvement approach where models iteratively generate and learn from their own solutions, progressively tackling harder problems while maintaining a standard transformer architecture. Across diverse tasks including arithmetic, string manipulation, and maze solving, self-improving enables models to solve problems far beyond their initial training distribution-for instance, generalizing from 10-digit to 100-digit addition without apparent saturation. We observe that in some cases filtering for correct self-generated examples leads to exponential improvements in out-of-distribution performance across training rounds. Additionally, starting from pretrained models significantly accelerates this self-improvement process for several tasks. Our results demonstrate how controlled weak-to-strong curricula can systematically teach a model logical extrapolation without any changes to the positional embeddings, or the model architecture.
Forward citations
Cited by 5 Pith papers
-
Symmetry-Aware Transformer Training for Automated Planning
A contrastive loss that aligns attention and hidden states between renamed copies helps transformers solve larger planning problems in some domains, but not all.
-
A Task-Centric Theory for Iterative Self-Improvement with Easy-to-Hard Curricula
Iterative self-improvement provably keeps improving only when initial performance lies in a moderate difficulty interval, and easy-to-hard curricula beat fixed mixtures under moderate difficulty separation and suffici...
-
Physics of Language Models: Part 4.1, Architecture Design and the Magic of Canon Layers
Canon layers—residual 1-d causal convolutions over adjacent tokens—boost synthetic reasoning depth 2-4x, lift NoPE to RoPE level, and bring GLA up to Mamba2/GDN, with qualitative real-world confirmation.
-
Maximizing Prefix-Confidence at Test-Time Efficiently Improves Mathematical Reasoning
Selecting the most confident 32-token prefix and completing only it gives better accuracy per compute than majority voting on five math reasoning datasets, using only the model's own confidence as a selector.
-
Extrapolation by Association: Length Generalization Transfer in Transformers
Length generalization on a short-trained main task can be inherited from a longer-trained related auxiliary task trained jointly with it.
Discussion (0). Sign in to comment.