princeton-nlp/LESS

• highreadme#1

  Clarify the repo's purpose as a data selection tool for LLM instruction tuning in the README's opening

  原因：

  当前

  This repo contains the code for our ICML 2024 paper LESS: Selecting Influential Data for Targeted Instruction Tuning. In this work, we propose a data selection method to select influential data to induce a target capability.

  复制粘贴的修复

  复制

  LESS provides a practical, code-based method for **selecting influential training data to enhance targeted instruction tuning of large language models (LLMs)**. This repository contains the official implementation for our ICML 2024 paper, 'LESS: Selecting Influential Data for Targeted Instruction Tuning,' which proposes this novel data selection approach to induce specific target capabilities.

  待办完成忽略
• mediumtopics#2

  Add more specific topics related to LLM fine-tuning and data optimization

  原因：

  当前

  data, data-selection, influence, instruction-tuning, llama, llm, mistral

  复制粘贴的修复

  复制

  data, data-selection, influence, instruction-tuning, llama, llm, mistral, llm-fine-tuning, data-optimization, model-tuning

  待办完成忽略
• lowhomepage#3

  Add the paper's arXiv link as the repository homepage

  原因：

  复制粘贴的修复

  复制

  https://arxiv.org/abs/2402.06020

  待办完成忽略

本次扫描解析到的品类 GEO 通道：google/gemini-2.5-flash, deepseek/deepseek-v4-flash

• 品类问题

  How can I select the most effective training data for targeted LLM instruction tuning?

  你：未被推荐
  AI 推荐顺序：

  1. Argilla (argilla-io/argilla)
  2. Label Studio (heartexlabs/label-studio)
  3. Snorkel (snorkel-team/snorkel)
  4. Cleanlab (cleanlab/cleanlab)
  5. OpenAI's API
  6. Claude
  7. Gemini
  8. Cohere's Embeddings
  9. OpenAI's `text-embedding-ada-002`
  10. GPT-4
  11. Claude 3
  12. Llama 3
  13. NLPAug (makcedward/nlpaug)
  14. TextAttack (textattack/textattack)
  15. Alpaca (tatsu-lab/stanford_alpaca)
  16. Dolly 2.0 (databrickslabs/dolly)
  17. ShareGPT
  18. FLAN

  AI 推荐了 18 个替代方案，却始终没点名 princeton-nlp/LESS。这就是要补上的差距。

  查看 AI 完整回答
• 品类问题

  What methods exist for identifying influential data points to enhance large language model fine-tuning?

  你：未被推荐
  AI 推荐顺序：

  1. PyTorch-Influence-Functions
  2. Shapley
  3. PyTorch
  4. TensorFlow
  5. Hugging Face Transformers
  6. modAL
  7. UMAP
  8. t-SNE
  9. PCA

  AI 推荐了 9 个替代方案，却始终没点名 princeton-nlp/LESS。这就是要补上的差距。

  查看 AI 完整回答

• Metadata completeness

  warn

  建议：

• README presence

  pass

• Compared to common alternatives in this category, what is the core differentiator of princeton-nlp/LESS?

  pass
  ✓AI 明确点名了 princeton-nlp/LESS

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• If a team adopts princeton-nlp/LESS in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI 明确点名了 princeton-nlp/LESS

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？

• In one sentence, what problem does the repo princeton-nlp/LESS solve, and who is the primary audience?

  pass
  ✓AI 明确点名了 princeton-nlp/LESS

  AI 的回答可能信誓旦旦却是错的。请按事实核对：技术栈、目标人群、差异化点是不是和你实际的对得上？