Pith. sign in

REVIEW 7 cited by

Entropy Law: The Story Behind Data Compression and LLM Performance

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

arxiv 2407.06645 v3 pith:UPMMNFR4 submitted 2024-07-09 cs.LG cs.CL

classification cs.LGcs.CL
keywords datacompressionentropyllmsperformancetrainingmodelratio
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
read the original abstract

Data is the cornerstone of large language models (LLMs), but not all data is useful for model learning. Carefully selected data can better elicit the capabilities of LLMs with much less computational overhead. Most methods concentrate on evaluating the quality of individual samples in data selection, while the combinatorial effects among samples are neglected. Even if each sample is of perfect quality, their combinations may be suboptimal in teaching LLMs due to their intrinsic homogeneity or contradiction. In this paper, we aim to uncover the underlying relationships between LLM performance and data selection. Inspired by the information compression nature of LLMs, we uncover an ``entropy law'' that connects LLM performance with data compression ratio and first-epoch training loss, which reflect the information redundancy of a dataset and the mastery of inherent knowledge encoded in this dataset, respectively. Through both theoretical deduction and empirical evaluation, we find that model performance is negatively correlated to the compression ratio of training data, which usually yields a lower training loss. Based on the findings of the entropy law, we propose a quite efficient and universal data selection method named \textbf{ZIP} for training LLMs, which aim to prioritize data subsets exhibiting a low compression ratio. Based on a multi-stage algorithm that selects diverse data in a greedy manner, we can obtain a good data subset with satisfactory diversity. Extensive experiments have been conducted to validate the entropy law and the superiority of ZIP across different LLM backbones and alignment stages. We also present an interesting application of entropy law that can detect potential performance risks at the beginning of model training.

Discussion (0). Continue with ORCID to comment.

Forward citations

Cited by 7 Pith papers

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. SelfAug: Mitigating Catastrophic Forgetting in Retrieval-Augmented Generation via Distribution Self-Alignment

    cs.CL 2025-09 conditional novelty 5.0 of 10

    Adding a KL penalty between fine-tuned and original model logits on input tokens during RAG fine-tuning reduces catastrophic forgetting while preserving downstream performance.

  2. Pangu Pro MoE: Mixture of Grouped Experts for Efficient Sparsity

    cs.CL 2025-05 conditional novelty 5.0 of 10

    A group-constrained Mixture-of-Experts routing rule (MoGE) is proposed and demonstrated in Pangu Pro MoE, a 72B/16B-active sparse LLM that reports faster inference on Ascend NPUs.

  3. DLF: Enhancing Explicit-Implicit Interaction via Dynamic Low-Order-Aware Fusion for CTR Prediction

    cs.IR 2025-05 conditional novelty 5.0 of 10

    DLF is a CTR prediction architecture that combines low-rank, high-rank, and implicit interaction blocks with layer-wise attention fusion, reporting state-of-the-art results on Criteo, Avazu, Movielens, and Frappe.

  4. FuXi-$\alpha$: Scaling Recommendation Model with Feature Interaction Enhanced Transformer

    cs.IR 2025-02 conditional novelty 5.0 of 10

    FuXi-alpha, a sequential recommender with decoupled temporal, positional, and semantic attention channels plus a two-stage FFN, reports gains over HSTU and positive online engagement results.

  5. Pangu Embedded: An Efficient Dual-system LLM Reasoner with Metacognition

    cs.CL 2025-05 conditional novelty 4.0 of 10

    Pangu Embedded, a 7B reasoner trained with iterative distillation, RL, and an adaptive fast/slow thinking scheme, reports superior benchmark scores to similarly sized Qwen3-8B and GLM-4-9B.

  6. The Paradox of Stochasticity: Limited Creativity and Computational Decoupling in Temperature-Varied LLM Outputs of Structured Fictional Data

    cs.LG 2025-02 conditional novelty 4.0 of 10

    In three LLMs generating fictional names and birthdates, model choice dominates processing time and default name archetypes persist across temperature, while rare names appear mainly at mid-range temperatures.

  7. Alignment and Safety in Large Language Models: Safety Mechanisms, Training Paradigms, and Emerging Challenges

    cs.AI 2025-07 reject novelty 1.0 of 10

    A broad survey of LLM alignment that catalogs objectives, benchmarks, SFT/RLHF/DPO methods, and safety challenges, without contributing new experimental or theoretical results.

Pith tools