Understanding Evolution Strategies for LLM Reasoning: Broader Reasoning Coverage than GRPO

Hugging Face Daily Papers Papers

Summary

This paper investigates Evolution Strategies (ES) as a post-training paradigm for LLM reasoning, showing that ES provides broader reasoning coverage and better Pass@K performance than GRPO through sparse functional updates and population diversity.

Evolution Strategies (ES) have recently emerged as a memory-efficient post-training paradigm for LLM reasoning. However, the optimization behavior of ES remains understudied, making it hard to define its advantage scope compared to mainstream post-training paradigms (e.g., Group Relative Policy Optimization (GRPO)). By systematically investigating ES dynamics and mechanisms, this paper first identifies a performance advantage of ES over GRPO, theoretically and empirically showing that ES can lead to broader reasoning coverage, thereby better exploiting the reasoning capabilities of pretrained LLMs. Theoretically, we show that verifier-projected Jensen-Shannon diversity across the ES population is helpful to higher Pass@K performances. Empirically, unlike GRPO, which exhibits entropy collapse, ES improves Pass@1 while attaining higher Pass@K than GRPO. We further develop a sequential GRPO-ES training strategy that combines GRPO's strength in Pass@1 with ES's gains in Pass@K. Second, we find that despite substantial whole-model parameter drift, the task-performance gains of ES are only contributed to a sparse subset of larger-magnitude updates. This functional sparsity suggests that large parameter movement need not imply widespread functional change, and held-out evaluations further show that it does not necessarily lead to catastrophic forgetting. Finally, we study how hyperparameter design affects the effectiveness of ES, demonstrating that ES requires a smaller population size in a larger LLM. These findings position ES as a distinct reasoning post-training paradigm rather than a less effective, memory-efficient alternative to GRPO.
Original Article
View Cached Full Text

Cached at: 08/28/26, 03:24 AM

Paper page - Understanding Evolution Strategies for LLM Reasoning: Broader Reasoning Coverage than GRPO

Source: https://huggingface.co/papers/2608.27351

Abstract

Evolution strategies improve reasoning diversity and Pass@K over GRPO through sparse functional updates and population diversity, supporting a hybrid training approach.

Evolution Strategies(ES) have recently emerged as a memory-efficient post-training paradigm for LLM reasoning. However, the optimization behavior of ES remains understudied, making it hard to define its advantage scope compared to mainstream post-training paradigms (e.g.,Group Relative Policy Optimization(GRPO)). By systematically investigating ES dynamics and mechanisms, this paper first identifies a performance advantage of ES overGRPO, theoretically and empirically showing that ES can lead to broader reasoning coverage, thereby better exploiting the reasoning capabilities of pretrained LLMs. Theoretically, we show thatverifier-projected Jensen-Shannon diversityacross the ES population is helpful to higherPass@Kperformances. Empirically, unlikeGRPO, which exhibitsentropy collapse, ES improves Pass@1 while attaining higherPass@KthanGRPO. We further develop a sequentialGRPO-ES training strategy that combinesGRPO’s strength in Pass@1 with ES’s gains inPass@K. Second, we find that despite substantial whole-model parameter drift, the task-performance gains of ES are only contributed to a sparse subset of larger-magnitude updates. Thisfunctional sparsitysuggests that large parameter movement need not imply widespread functional change, and held-out evaluations further show that it does not necessarily lead tocatastrophic forgetting. Finally, we study how hyperparameter design affects the effectiveness of ES, demonstrating that ES requires a smallerpopulation sizein a larger LLM. These findings position ES as a distinct reasoning post-training paradigm rather than a less effective, memory-efficient alternative toGRPO.

View arXiv pageView PDFAdd to collection

Get this paper in your agent:

hf papers read 2608\.27351

Don’t have the latest CLI?curl \-LsSf https://hf\.co/cli/install\.sh \| bash

Models citing this paper0

No model linking this paper

Cite arxiv.org/abs/2608.27351 in a model README.md to link it from this page.

Datasets citing this paper0

No dataset linking this paper

Cite arxiv.org/abs/2608.27351 in a dataset README.md to link it from this page.

Spaces citing this paper0

No Space linking this paper

Cite arxiv.org/abs/2608.27351 in a Space README.md to link it from this page.

Collections including this paper0

No Collection including this paper

Add this paper to acollectionto link it from this page.

Similar Articles

When Do LLMs Reason? A Dynamical Systems View via Entropy Phase Transitions

arXiv cs.LG

This paper investigates when chain-of-thought reasoning is beneficial for LLMs, showing that early-stage entropy dynamics reliably indicate reasoning utility, and introduces EDRM, a lightweight, training-free framework that adaptively selects inference strategies to achieve significant token savings while maintaining or improving accuracy.