ESSA: Evolutionary Strategies for Scalable Alignment

• URL: http://arxiv.org/abs/2507.04453v2
• Date: Sat, 18 Oct 2025 11:12:14 GMT
• Title: ESSA: Evolutionary Strategies for Scalable Alignment
• Authors: Daria Korotyshova, Boris Shaposhnikov, Alexey Malakhov, Alexey Khokhulin, Nikita Surnachev, Kirill Ovcharenko, George Bredis, Alexey Gorbatovski, Viacheslav Sinii, Daniil Gavrilov,
• Abstract summary: We present ESSA, a gradient-free framework that aligns Large Language Models (LLMs) using only forward inference and black-box optimization.<n>ESSA improves the test accuracy of Qwen2.5-Math-7B by 12.6% on GSM8K and 14.8% on PRM800K, and raises the accuracy of LLaMA3.1-8B on IFEval by 22.5%.<n>In large-scale settings ESSA shows stronger scaling than gradient-based methods.
• Score: 8.418036456622158
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Alignment of Large Language Models (LLMs) typically relies on Reinforcement Learning from Human Feedback (RLHF) with gradient-based optimizers such as Proximal Policy Optimization (PPO) or Group Relative Policy Optimization (GRPO). While effective, these methods require complex distributed training, large memory budgets, and careful hyperparameter tuning, all of which become increasingly difficult at billion-parameter scale. We present ESSA, Evolutionary Strategies for Scalable Alignment, a gradient-free framework that aligns LLMs using only forward inference and black-box optimization. ESSA focuses optimization on Low-Rank Adapters (LoRA) and further compresses their parameter space by optimizing only the singular values from an SVD decomposition of each adapter matrix. This dimensionality reduction makes evolutionary search practical even for very large models and allows efficient operation in quantized INT4 and INT8 inference mode. Across these benchmarks ESSA improves the test accuracy of Qwen2.5-Math-7B by 12.6% on GSM8K and 14.8% on PRM800K, and raises the accuracy of LLaMA3.1-8B on IFEval by 22.5%, all compared with GRPO. In large-scale settings ESSA shows stronger scaling than gradient-based methods: on Qwen2.5-32B for PRM800K it reaches near-optimal accuracy twice as fast on 16 GPUs and six times as fast on 128 GPUs compared with GRPO. These results position evolutionary strategies as a compelling, hardware-friendly alternative to gradient-based LLM alignment, combining competitive quality with substantially reduced wall-clock time and engineering overhead.

Related papers

• Beyond KL Divergence: Policy Optimization with Flexible Bregman Divergences for LLM Reasoning [3.259050650999544]
  Group-Based Mirror Policy Optimization (GBMPO) is a framework that extends group-based policy optimization to flexible Bregman divergences.<n>Hand-designed ProbL2-GRPO achieves 86.7% accuracy, improving +5.5 points over the Dr. GRPO baseline.
  arXiv  Detail & Related papers  (2026-02-04T10:01:20Z)
• Hyperparameter Transfer Enables Consistent Gains of Matrix-Preconditioned Optimizers Across Scales [55.91454326946738]
  We study how the optimal learning rate and weight decay should scale with model width and depth for a wide range of languages.<n>We find that scaling the learning rate according to $$P improves transfer, but can still suffer from significant finite-width deviations.<n>For compute-optimal scaling, we find scaling independent weight decay as $1/mathrmwidth$ is nearly optimal across languages.
  arXiv  Detail & Related papers  (2025-12-05T11:03:41Z)
• Tri-Accel: Curvature-Aware Precision-Adaptive and Memory-Elastic Optimization for Efficient GPU Usage [0.6511750267058007]
  Tri-Accel is a unified optimization framework that co-adapts three acceleration strategies along with adaptive parameters during training.<n>On CIFAR-10 with ResNet-18 and EfficientNet-B0, Tri-Accel achieves up to 9.9% reduction in training time and 13.3% lower memory usage.<n>Compared to static mixed-precision training, Tri-Accel maintains 78.1% accuracy while reducing memory footprint from 0.35GB to 0.31GB on standard hardware.
  arXiv  Detail & Related papers  (2025-08-23T05:38:42Z)
• Hybrid and Unitary Fine-Tuning of Large Language Models: Methods and Benchmarking under Resource Constraints [0.0]
  Fine-tuning large language models (LLMs) remains a computational bottleneck due to their scale and memory demands.<n>This paper presents a comprehensive evaluation of parameter-efficient fine-tuning (PEFT) techniques, including LoRA, BOFT, LoRA-GA, and uRNN.<n>By computing per-layer adaptive updates guided by gradient norms, the hybrid method achieves superior convergence efficiency and generalization across diverse tasks.
  arXiv  Detail & Related papers  (2025-07-24T04:00:02Z)
• Beyond First-Order: Training LLMs with Stochastic Conjugate Subgradients and AdamW [2.028622227373579]
  gradient-based descent (SGD) have long been central to training large language models (LLMs)<n>This paper proposes a conjugate subgradient method together with adaptive sampling specifically for training LLMs.
  arXiv  Detail & Related papers  (2025-07-01T23:30:15Z)
• Taming LLMs by Scaling Learning Rates with Gradient Grouping [49.91587150497186]
  Training large language models (LLMs) poses challenges due to their massive scale and heterogeneous architectures.<n>This work introduces Scaling with Gradient Grouping (SGG), an gradient wrapper that improves adaptive learning rate estimation by dynamic grouping and group-specific scaling.
  arXiv  Detail & Related papers  (2025-06-01T15:30:37Z)
• LLaDA 1.5: Variance-Reduced Preference Optimization for Large Language Diffusion Models [95.77351099118323]
  Masked Diffusion Models (MDMs) present a promising paradigm for language modeling.<n>The challenge arises from the high variance in Evidence Lower Bound (ELBO)-based likelihood estimates required for preference optimization.<n>We propose Variance-Reduced Preference Optimization (VRPO), a framework that formally analyzes the variance of ELBO estimators and derives on both the bias and variance of preference optimization gradients.
  arXiv  Detail & Related papers  (2025-05-25T16:36:20Z)
• KerZOO: Kernel Function Informed Zeroth-Order Optimization for Accurate and Accelerated LLM Fine-Tuning [15.81250204481401]
  We introduce a kernel-function-based ZO framework aimed at mitigating gradient estimation bias.<n>KerZOO achieves comparable or superior performance to existing ZO baselines.<n>We show that the kernel function is an effective avenue for reducing estimation bias in ZO methods.
  arXiv  Detail & Related papers  (2025-05-24T21:56:03Z)
• AYLA: Amplifying Gradient Sensitivity via Loss Transformation in Non-Convex Optimization [0.0]
  Gradient Descent (SGD) and its variants, such as ADAM, are foundational to deep learning optimization.<n>This paper introduces AYLA, a novel framework that enhances dynamics training.
  arXiv  Detail & Related papers  (2025-04-02T16:31:39Z)
• LORENZA: Enhancing Generalization in Low-Rank Gradient LLM Training via Efficient Zeroth-Order Adaptive SAM [13.180761892449736]
  We study robust parameter-efficient fine-tuning (PEFT) techniques for large-language models (LLMs)<n>We present a new highly computationally efficient framework called AdaZo-SAM, combining Adam and Sharpness-Aware Minimization (SAM)<n>We also design a low-rank gradient optimization method named LORENZA, which is a memory-efficient version of AdaZo-SAM.
  arXiv  Detail & Related papers  (2025-02-26T21:30:34Z)
• COSMOS: A Hybrid Adaptive Optimizer for Memory-Efficient Training of LLMs [81.01082659623552]
  Large Language Models (LLMs) have demonstrated remarkable success across various domains.<n>Their optimization remains a significant challenge due to the complex and high-dimensional loss landscapes they inhabit.
  arXiv  Detail & Related papers  (2025-02-24T18:42:19Z)
• Gradient Multi-Normalization for Stateless and Scalable LLM Training [16.037614012166063]
  Training large language models (LLMs) typically relies on adaptives like Adam which store additional state information to accelerate convergence but incur significant memory overhead.<n>Recent efforts, such as SWAN (Ma et al., 2024) address this by eliminating the need for states while achieving performance comparable to Adam via a multi-step preprocessing procedure applied to instantaneous gradients.<n>We introduce a novel framework for designing stateless gradients that normalizes gradients according to multiple norms. Experiments on pre-training LLaMA models with up to 1 billion parameters demonstrate a 3X speedup over Adam with significantly reduced memory requirements, outperforming other memory-efficient baseline
  arXiv  Detail & Related papers  (2025-02-10T18:09:53Z)
• HALO: Hadamard-Assisted Lower-Precision Optimization for LLMs [48.55966021231297]
  We present HALO, a novel quantization-aware training approach for Transformers.<n>Our approach ensures that all large matrix multiplications during the forward and backward passes are executed in lower precision.<n>Applying to LLAMA-family models, HALO achieves near-full-precision-equivalent results during fine-tuning on various tasks.
  arXiv  Detail & Related papers  (2025-01-05T18:41:54Z)
• Constrain Alignment with Sparse Autoencoders [47.15096507230884]
  Feature-level constrained Preference Optimization is a novel method designed to simplify the alignment process while ensuring stability.<n>Our approach enjoys efficiency by using sparse features activated in a well-trained sparse autoencoder and the quality of sequential KL divergence.
  arXiv  Detail & Related papers  (2024-11-12T07:54:13Z)
• HAFLQ: Heterogeneous Adaptive Federated LoRA Fine-tuned LLM with Quantization [55.972018549438964]
  Federated fine-tuning of pre-trained Large Language Models (LLMs) enables task-specific adaptation across diverse datasets while preserving privacy.<n>We propose HAFLQ (Heterogeneous Adaptive Federated Low-Rank Adaptation Fine-tuned LLM with Quantization), a novel framework for efficient and scalable fine-tuning of LLMs in heterogeneous environments.<n> Experimental results on the text classification task demonstrate that HAFLQ reduces memory usage by 31%, lowers communication cost by 49%, improves accuracy by 50%, and achieves faster convergence compared to the baseline method.
  arXiv  Detail & Related papers  (2024-11-10T19:59:54Z)
• Towards Explainable Evolution Strategies with Large Language Models [0.0]
  This paper introduces an approach that integrates self-adaptive Evolution Strategies (ES) with Large Language Models (LLMs) By employing a self-adaptive ES equipped with a restart mechanism, we effectively navigate the challenging landscapes of benchmark functions. An LLM is then utilized to process these logs, generating concise, user-friendly summaries.
  arXiv  Detail & Related papers  (2024-07-11T09:28:27Z)
• Adaptive Preference Scaling for Reinforcement Learning with Human Feedback [103.36048042664768]
  Reinforcement learning from human feedback (RLHF) is a prevalent approach to align AI systems with human values. We propose a novel adaptive preference loss, underpinned by distributionally robust optimization (DRO) Our method is versatile and can be readily adapted to various preference optimization frameworks.
  arXiv  Detail & Related papers  (2024-06-04T20:33:22Z)
• AffineQuant: Affine Transformation Quantization for Large Language Models [58.45460102764]
  Post-Training Quantization (PTQ) has emerged as a subject of considerable interest due to its compression efficiency and cost-effectiveness in the context of training. Existing PTQ methods for Large-scale Language Models (LLMs) limit the optimization scope to scaling transformations between pre- and post-quantization weights. In this paper, we advocate for the direct optimization using equivalent Affine transformations in PTQ (AffineQuant)
  arXiv  Detail & Related papers  (2024-03-19T08:40:21Z)
• Revisiting Zeroth-Order Optimization for Memory-Efficient LLM Fine-Tuning: A Benchmark [166.40879020706151]
  This paper proposes a shift towards BP-free, zeroth-order (ZO) optimization as a solution for reducing memory costs during fine-tuning. Unlike traditional ZO-SGD methods, our work expands the exploration to a wider array of ZO optimization techniques. Our study unveils previously overlooked optimization principles, highlighting the importance of task alignment, the role of the forward gradient method, and the balance between algorithm complexity and fine-tuning performance.
  arXiv  Detail & Related papers  (2024-02-18T14:08:48Z)
• AdaLomo: Low-memory Optimization with Adaptive Learning Rate [59.64965955386855]
  We introduce low-memory optimization with adaptive learning rate (AdaLomo) for large language models. AdaLomo results on par with AdamW, while significantly reducing memory requirements, thereby lowering the hardware barrier to training large language models.
  arXiv  Detail & Related papers  (2023-10-16T09:04:28Z)
• Evolving Pareto-Optimal Actor-Critic Algorithms for Generalizability and Stability [67.8426046908398]
  Generalizability and stability are two key objectives for operating reinforcement learning (RL) agents in the real world. This paper presents MetaPG, an evolutionary method for automated design of actor-critic loss functions.
  arXiv  Detail & Related papers  (2022-04-08T20:46:16Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.