Logic-Induced-Long-Tail (LINT)

Huihan Li 0001, Yuting Ning, Zeyi Liao, Xiang Ren 0001, Huihan Li, Xiang Ren2023 · DOI

Introduction: The Challenge of the Long-Tail

Large Language Models (LLMs) are primarily trained on the "head" of the data distribution—the most frequent facts and concepts found across the internet. While these models demonstrate remarkable proficiency in common scenarios, they often falter when encountering "long-tail" knowledge: rare, specific, or unfamiliar facts where model confidence is naturally lower. Evaluating this capability is difficult due to data contamination, human cognitive biases toward common examples, and the inherent tendency of LLMs to predict the most likely next token.

To address these challenges, researchers introduced LINK (Logic-Induced-Knowledge-Search). Unlike standard prompting, LINK utilizes a structured, "variable-wise" search guided by symbolic logic templates to ensure that the generated knowledge is both rare and factually sound.

The LINK framework for generating long-tail knowledge via symbolic rules, verification, and reranking.

The framework operates in four primary stages:

  1. Symbolic Rule Definition: Logical templates (e.g., temporal constraints) are established to ensure deductive soundness.
  2. Variable-Wise Prompting: The model fills specific variables within the template sequentially rather than generating a full sentence at once.
  3. Verification: A "critic" model (Flan-T5-XXL) verifies factual correctness and data type conformity.
  4. Knowledge Beam Search & Reranking: A smaller model (Llama-7B) calculates the likelihood of generated statements, allowing the system to deliberately select low-probability "long-tail" values.

The LINT Dataset

The result of this process is LINT (Logic-Induced-Long-Tail), a comprehensive dataset of 108,000 inferential statements. LINT spans four critical domains: Temporal, Locational, Outcome and Effect, and Natural Properties. Each entry is structured as a Natural Language Inference (NLI) task, providing a rigorous benchmark for testing whether LLMs can reason through rare information or are merely relying on memorized patterns from their training data.

Results: LLM Performance on Rare Knowledge

Evaluation across several state-of-the-art models revealed a significant performance gap between common "head" knowledge and rare "long-tail" knowledge.

Comparative performance of Llama2, ChatGPT, and GPT-4 on head vs. long-tail datasets, showing a significant drop for all models.

Key findings include:

  • Stark Performance Drops: Even the most advanced models showed vulnerability. GPT-4 experienced a 21% relative drop in accuracy on the long-tail set. Open-source and earlier proprietary models like Llama2-70B and ChatGPT saw performance collapses exceeding 50%.
  • Reasoning Brittleness: Models often changed their answers when the same logical fact was presented with slight variations in phrasing, suggesting a lack of robust underlying reasoning.
  • Human Stability: In contrast to LLMs, human performance remained stable across both common and rare facts (when provided with search tools), highlighting a fundamental difference in how humans and machines generalize knowledge.

Conclusion

The LINT dataset and LINK framework demonstrate that current LLMs are heavily reliant on the frequency of data in their training sets. By grounding generation in symbolic logic, researchers can systematically identify these "blind spots." Future work will look toward expanding these methods to more complex logical structures and reducing the computational overhead of the search process.

Summary updated

BibTeX
@misc{li2023logic,
  author = {Huihan Li and Yuting Ning and Zeyi Liao and Xiang Ren and Huihan Li and Xiang Ren and Yuting Ning and Zeyi Liao},
  title = {{Logic-Induced-Long-Tail (LINT)}},
  year = {2023},
  doi = {10.5281/zenodo.10117911},
}

Loading related content...

Figure