# LASER: Low-Rank Activation SVD for Efficient Recursion

arXiv: https://arxiv.org/abs/2604.17224 · HTML: https://mlsanity.com/paper/2604.17224/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning.
- Builds on 5 of its 10 references that are in the map.
- First posted on arXiv 19 Apr 2026.

## Placed in

- LLM reasoning / [Test-time compute and efficient thinking](https://mlsanity.com/lr/test-time-compute-and-efficient-thinking/)

## Builds on (most cited first, 5 of 5)

- [The Unreasonable Ineffectiveness of the Deeper Layers](https://mlsanity.com/paper/2403.17887/) (2024)
- [Scaling Latent Reasoning via Looped Language Models](https://mlsanity.com/paper/2510.25741/) (2025)
- [ActNN: Reducing Training Memory Footprint via 2-Bit Activation Compressed Training](https://mlsanity.com/paper/2104.14129/) (2021)
- [LANCE: Low Rank Activation Compression for Efficient On-Device Continual Learning](https://mlsanity.com/paper/2509.21617/) (2025)
- [Memory-Efficient Fine-Tuning via Low-Rank Activation Compression](https://mlsanity.com/paper/2509.23472/) (2025)

## Built on by (most cited first, 1 of 1)

- [Loop the Loopies!](https://mlsanity.com/paper/2607.16051/) (2026)

## Abstract (arXiv)

Recursive architectures such as Tiny Recursive Models (TRMs) perform implicit reasoning through iterative latent computation, yet the geometric structure of these reasoning trajectories remains poorly understood. We investigate the activation manifold of TRMs during recursive unrolling and find that activations occupy an effectively linear, low-dimensional subspace whose principal directions can be tracked dynamically with cheap power iterations. This suggests that weight-sharing concentrates iterative computation along a small number of dominant eigendirections, and we find that this concentration varies sharply across computational sites. We exploit this structure through LASER (Low-Rank Activation SVD for Efficient Recursion), a dynamic compression framework that maintains an evolving low-rank basis via matrix-free subspace tracking with a fidelity-triggered reset mechanism, achieving ~60% activation memory savings with no statistically significant accuracy degradation. Our analysis raises questions about how recursive architectures allocate representational capacity during implicit reasoning, and whether this concentration can be exploited to improve the efficiency and stability of latent computation.