Summary of Sled: Self Logits Evolution Decoding For Improving Factuality in Large Language Models, by Jianyi Zhang et al.
SLED: Self Logits Evolution Decoding for Improving Factuality in Large Language Models
by Jianyi Zhang, Da-Cheng Juan, Cyrus Rashtchian, Chun-Sung Ferng, Heinrich Jiang, Yiran Chen
First submitted to arxiv on: 1 Nov 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (stat.ML)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary The paper introduces Self Logits Evolution Decoding (SLED), a novel framework that enhances the truthfulness of Large Language Models (LLMs) without relying on external knowledge bases or requiring fine-tuning. SLED leverages latent knowledge within the LLM by contrasting output logits from the final layer with those from early layers, and utilizes an approximate gradient approach to guide self-refinement of outputs. This improves factual accuracy in tasks such as multi-choice, open-generation, and chain-of-thought reasoning. The results demonstrate up to 20% improvement in factual accuracy compared to existing decoding methods, while maintaining natural language fluency and negligible latency overhead. SLED can be flexibly combined with other decoding methods to further enhance performance. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary The paper talks about how big AI models called Large Language Models (LLMs) sometimes give wrong answers or facts. To fix this, the authors created a new way of making these models better called Self Logits Evolution Decoding (SLED). SLED helps make the model’s answers more accurate by looking at what it knows already and using that to improve its answers. The authors tested this method on many different AI models and showed that it makes their answers about 20% more accurate, without making them sound unnatural or taking a long time. |
Keywords
» Artificial intelligence » Fine tuning » Logits
