inverse-scaling/prize

• highreadme#1

  Reposition README opening to clarify archive status

  原因：

  当前

  Inverse Scaling Prize
  **TL;DR: Win up to $100,000 for finding an important task where larger language models do worse.**
  _~~Submissions due August 27, 2022 (Round 1) and October 27, 2022 (Round 2).~_
  The contest has ended! Results: Round 1, Round 2.

  复制粘贴的修复

  复制

  Inverse Scaling Prize: Archive of Past Results and Datasets
  
  This repository serves as the official archive for the Inverse Scaling Prize, a concluded competition focused on identifying tasks where larger language models perform worse. It provides access to the winning tasks, datasets, and detailed results from Round 1 and Round 2, offering a valuable resource for researchers studying inverse scaling phenomena in LLMs.

  待办完成忽略
• hightopics#2

  Add relevant topics for categorization

  原因：

  复制粘贴的修复

  复制

  large-language-models, llm-evaluation, inverse-scaling, ai-safety, machine-learning-datasets, research-archive, benchmark-datasets

  待办完成忽略
• mediumabout#3

  Update repository description to reflect archive status

  原因：

  当前

  A prize for finding tasks that cause large language models to show inverse scaling

  复制粘贴的修复

  复制

  Archive of a past prize for finding tasks that cause large language models to show inverse scaling, including results and datasets.

  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 to identify scenarios where larger language models exhibit unexpected performance degradation?

  你：未被推荐
  AI 推荐顺序：

  1. TextAttack (TextAttack/TextAttack)
  2. OpenAttack (THUDM/OpenAttack)
  3. LIME (marcotcr/lime)
  4. SHAP (shap/shap)
  5. Captum (pytorch/captum)

  AI 推荐了 5 个替代方案，却始终没点名 inverse-scaling/prize。这就是要补上的差距。

  查看 AI 完整回答
• 品类问题

  Where can I find datasets or benchmarks to test limitations of large language models?

  你：未被推荐
  AI 推荐顺序：

  1. Hugging Face Datasets Hub
  2. EleutherAI's LM Evaluation Harness
  3. BIG-bench
  4. MMLU
  5. HELM
  6. Adversarial NLI
  7. TruthfulQA

  AI 推荐了 7 个替代方案，却始终没点名 inverse-scaling/prize。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 inverse-scaling/prize?

  pass
  ✓AI 明确点名了 inverse-scaling/prize

  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 inverse-scaling/prize in production, what risks or prerequisites should they evaluate first?

  pass
  ✓AI 明确点名了 inverse-scaling/prize

  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 inverse-scaling/prize solve, and who is the primary audience?

  pass
  ⚠AI 未点名 inverse-scaling/prize —— 很可能在说另一个项目

  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？