Summary of Evaluating Zero-shot Long-context Llm Compression, by Chenyu Wang and Yihan Wang and Kai Li
Evaluating Zero-Shot Long-Context LLM Compression
by Chenyu Wang, Yihan Wang, Kai Li
First submitted to arxiv on: 10 Jun 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI)
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 evaluates the effectiveness of zero-shot compression techniques on large language models (LLMs) under long-context, identifying a trend for computational errors to increase when using certain methods. A hypothesis is proposed to explain the varied behavior of different LLM compression techniques and remedies are explored to mitigate performance decline in some techniques under long-context. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study looks at how well zero-shot compression works on big language models (LLMs) when they have a lot of context. It finds that some methods make mistakes as the context gets longer. The paper suggests why this might happen and tries to find ways to fix it so the LLMs don’t get worse under long-context. |
Keywords
» Artificial intelligence » Zero shot
