Summary of Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention, by Tsendsuren Munkhdalai et al.
Leave No Context Behind: Efficient Infinite Context Transformers with Infini-attention
by Tsendsuren Munkhdalai, Manaal Faruqui, Siddharth Gopal
First submitted to arxiv on: 10 Apr 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Machine Learning (cs.LG); Neural and Evolutionary Computing (cs.NE)
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 This paper proposes an efficient method to scale Transformer-based Large Language Models (LLMs) to infinitely long inputs, using a new attention technique called Infini-attention. The Infini-attention incorporates compressive memory into the vanilla attention mechanism, allowing for both masked local and long-term linear attention in a single Transformer block. The proposed approach is demonstrated on long-context language modeling benchmarks, including 1M sequence length passkey context block retrieval and 500K length book summarization tasks with 1B and 8B LLMs. |
Low | GrooveSquid.com (original content) | Low Difficulty Summary This paper finds a way to make very big language models work really well even when they’re faced with super long texts. They do this by inventing a new attention technique called Infini-attention, which is like a special kind of memory that helps the model focus on what’s important. This means that these language models can now handle tasks like summarizing entire books or searching for specific information in really long passages. |
Keywords
» Artificial intelligence » Attention » Summarization » Transformer