Pith. sign in

REVIEW 12 cited by

AlpaGasus: Training A Better Alpaca with Fewer Data

Not yet reviewed by Pith; the record is open.

Re-run · record.json Download PDF Read on arXiv ↗

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 2307.08701 v5 pith:QT7ZZCQ3 submitted 2023-07-17 cs.CL

Lichang Chen , Shiyang Li , Jun Yan , Hai Wang , Kalpa Gunaratna , Vikas Yadav , Zheng Tang , Vijay Srinivasan
show 3 more authors
Tianyi Zhou Heng Huang Hongxia Jin
This is my paper
classification cs.CL
keywords dataalpacaalpagasustrainingmodelsbetterdatasetsfaster
verification ladder T0 review T1 audit T2 compute T3 formal
0 comments
abstract

Large language models (LLMs) strengthen instruction-following capability through instruction-finetuning (IFT) on supervised instruction/response data. However, widely used IFT datasets (e.g., Alpaca's 52k data) surprisingly contain many low-quality instances with incorrect or irrelevant responses, which are misleading and detrimental to IFT. In this paper, we propose a simple and effective data selection strategy that automatically identifies and filters out low-quality data using a strong LLM (e.g., ChatGPT). To this end, we introduce AlpaGasus, which is finetuned on only 9k high-quality data filtered from the 52k Alpaca data. AlpaGasus significantly outperforms the original Alpaca as evaluated by GPT-4 on multiple test sets and the controlled human evaluation. Its 13B variant matches $>90\%$ performance of its teacher LLM (i.e., Text-Davinci-003 generating the 52k data) on test tasks. It also provides 5.7x faster training, reducing the training time for a 7B variant from 80 minutes (for Alpaca) to 14 minutes. Moreover, the experiments prove the efficacy of our method across diverse datasets, base models, and LLM filters. Overall, AlpaGasus demonstrates a novel data-centric IFT paradigm that can be generally applied to instruction-tuning data, leading to faster training and better instruction-following models. Our project page is available at: https://lichang-chen.github.io/AlpaGasus/

Discussion (0). Sign in to comment.

Forward citations

Cited by 12 Pith papers

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

  1. Towards Efficient and Effective Alignment of Large Language Models

    cs.CL 2025-06 conditional novelty 7.0 of 10

    A thesis presenting Lion, WebR, LTE, BMC, and FollowBench, five empirical methods that together address LLM alignment data, training, and evaluation.

  2. SFGA: A Statistics-First Gating Architecture with Adjudicative Escalation for Trustworthy SFT Data Procurement

    cs.LG 2026-07 conditional novelty 6.0 of 10

    A cost-aware gate that routes simple SFT procurement cases to cheap statistics and hard cases to an adjudicated LLM debate reports 0.90 accuracy on a synthetic closed-loop benchmark.

  3. Resolving Knowledge Conflicts in Domain-specific Data Selection: A Case Study on Medical Instruction-tuning

    cs.CL 2025-05 conditional novelty 6.0 of 10

    KDS selects instruction-tuning data by measuring knowledge alignment (NLI entailment vs reference) and knowledge consistency (cluster entropy of sampled responses), and reports gains on medical QA benchmarks.

  4. Merge to Mix: Mixing Datasets via Model Merging

    cs.LG 2025-05 conditional novelty 6.0 of 10

    Merge to Mix shows that the performance of a parameter-averaged model predicts the performance of a model fine-tuned on any dataset mixture, enabling fast and accurate dataset mixture selection.

  5. Less Data, Faster Convergence: Goal-Driven Data Optimization for Multimodal Instruction Tuning

    cs.CV 2026-03 conditional novelty 5.5 of 10

    Goal-driven selection of 1× multimodal instruction subsets reaches a 512k Uni-10x baseline after ~27–35k samples and improves accuracy by up to +3.08 pp under a fixed Qwen3-VL recipe.

  6. DaMoC: Efficiently Selecting the Optimal Large Language Model for Fine-tuning Domain Tasks Based on Data and Model Compression

    cs.CL 2025-09 reject novelty 5.0 of 10

    DaMoC combines data filtering, token compression, and layer pruning to select the best LLM for domain fine-tuning, claiming ~20x faster training while preserving model rankings.

  7. From Black Box to Transparency: Enhancing Automated Interpreting Assessment with Explainable AI in College Classrooms

    cs.CL 2025-08 unverdicted novelty 5.0 of 10

    An explainable model using BLEURT, CometKiwi, pause features, and Chinese phraseological diversity predicts human-rated quality dimensions in English-Chinese consecutive interpreting, with SHAP identifying the stronge...

  8. Efficient Data Selection at Scale via Influence Distillation

    cs.CL 2025-05 conditional novelty 5.0 of 10

    Influence Distillation selects LLM fine-tuning data by approximating each sample's gradient influence on a target task via landmarks and JVP embeddings, matching or beating RDS+ accuracy at roughly one third the selec...

  9. ALPS: Attention Localization and Pruning Strategy for Efficient Alignment of Large Language Models

    cs.CL 2025-05 conditional novelty 5.0 of 10

    ALPS selects task-sensitive attention heads by measuring Wasserstein distance between base and task-tuned weights, and freezing other heads during fine-tuning improves performance and efficiency.

  10. Hierarchical Fine-grained Preference Optimization for Physically Plausible Video Generation

    cs.CV 2025-08 conditional novelty 4.0 of 10

    A hierarchical direct preference optimization with four alignment levels plus automated data selection improves physical plausibility of text-to-video models.

  11. Improving Task Diversity in Label Efficient Supervised Finetuning of LLMs

    cs.CL 2025-07 conditional novelty 4.0 of 10

    Weighted Task Diversity allocates the annotation budget across tasks in inverse proportion to the base model's average confidence, improving MMLU and AlpacaEval scores with up to 80% fewer labels.

  12. Data Diversification Methods In Alignment Enhance Math Performance In LLMs

    cs.AI 2025-07 reject novelty 4.0 of 10

    DTS, which generates diverse solution strategies before writing solutions, improves GSM8K by 7.1 points and MATH by 4.2 points over an untuned base model at 1.03x baseline compute.

Pith tools