Papers for

model compression teams

Papers whose findings have a practical use for this group, as judged from the abstract. Open a paper to read what it means in practice.

Transformer updates separate direction and scale for robust learning

Disentangling Representation Evolution in Transformers through Directional Decomposition

Abstract: Transformer representations evolve through learned additive transformations that either preserve their current direction or redirect it. We study this evolution as a functional geometry, decomposing learned updates into parallel and perpendicular components. Across pretrained models, we find substantial parallel components beyond the residual identity path. We then apply the decomposition in two spaces: to attention and MLP updates relative to the hidden state, and to attention value aggregation relative to the current token's value. Targeted edits reveal a strongly space-dependent asymmetry: exclude-self value-space parallel manipulation is markedly more robust than residual-space and perpendicular counterparts, preserving the direct self message while scaling only the non-self aggregate. The same decomposition gives a component-resolved description of compression-induced update error: perpendicular error separates compression methods more clearly than parallel error. Extensive experiments further demonstrate that full-aggregate parallel suppression during from-scratch pretraining lowers validation-loss trajectories and improves downstream averages, with the value-space variant strongest. Together, these results connect representation geometry to editing robustness, compression diagnosis, and training-time intervention. Code is available in the \href{https://github.com/Shwai-He/Transformer-Geometry}{project repository}.

Mon 14 SeptComputation and LanguageMachine Learning
The gist
Transformers learn by changing their internal data representations step by step. The authors show that these changes can be split into two parts: one that keeps the same direction and one that changes it. They find that the part keeping the direction is surprisingly important and behaves differently in different parts of the model. By understanding this, they can make the model’s training more stable and fix errors better when compressing the model. This insight helps improve how transformers learn and are edited.
Open 2609.15975v1