# SirLLM: Streaming Infinite Retentive LLM

arXiv: https://arxiv.org/abs/2405.12528 · HTML: https://mlsanity.com/paper/2405.12528/

## Key facts

- Cited by 16 papers in the map (data through 14 Oct 2026).
- Builds on 22 of its 31 references that are in the map.
- First posted on arXiv 21 May 2024.

## Builds on (most cited first, 22 of 22)

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [GPT-4 Technical Report](https://mlsanity.com/paper/2303.08774/) (2023)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Llama 2: Open Foundation and Fine-Tuned Chat Models](https://mlsanity.com/paper/2307.09288/) (2023)
- [Longformer: The Long-Document Transformer](https://mlsanity.com/paper/2004.05150/) (2020)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [Efficient Streaming Language Models with Attention Sinks](https://mlsanity.com/paper/2309.17453/) (2023)
- [CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge](https://mlsanity.com/paper/1811.00937/) (2018)
- [Generating Long Sequences with Sparse Transformers](https://mlsanity.com/paper/1904.10509/) (2019)
- [Large Language Models](https://mlsanity.com/paper/2307.05782/) (2023)
- [DailyDialog: A Manually Labelled Multi-turn Dialogue Dataset](https://mlsanity.com/paper/1710.03957/) (2017)
- [Enhancing the Locality and Breaking the Memory Bottleneck of Transformer on Time Series Forecasting](https://mlsanity.com/paper/1907.00235/) (2019)
- [Model Tells You What to Discard: Adaptive KV Cache Compression for LLMs](https://mlsanity.com/paper/2310.01801/) (2023)
- [LongT5: Efficient Text-To-Text Transformer for Long Sequences](https://mlsanity.com/paper/2112.07916/) (2021)
- [CodeAid: Evaluating a Classroom Deployment of an LLM-based Programming Assistant that Balances Student and Educator Needs](https://mlsanity.com/paper/2401.11314/) (2024)
- [Advancing Transformer Architecture in Long-Context Large Language Models: A Comprehensive Survey](https://mlsanity.com/paper/2311.12351/) (2023)
- [CoCo-Agent: A Comprehensive Cognitive MLLM Agent for Smartphone GUI Automation](https://mlsanity.com/paper/2402.11941/) (2024)
- [Automating Customer Service using LangChain: Building custom open-source GPT Chatbot for organizations](https://mlsanity.com/paper/2310.05421/) (2023)
- [LLM as Prompter: Low-resource Inductive Reasoning on Arbitrary Knowledge Graphs](https://mlsanity.com/paper/2402.11804/) (2024)
- [Optimizing Retrieval-augmented Reader Models via Token Elimination](https://mlsanity.com/paper/2310.13682/) (2023)
- [ArcGPT: A Large Language Model Tailored for Real-world Archival Applications](https://mlsanity.com/paper/2307.14852/) (2023)
- [Approach Intelligent Writing Assistants Usability with Seven Stages of Action](https://mlsanity.com/paper/2304.02822/) (2023)

## Built on by (most cited first, 16 of 16)

- [Squeezed Attention: Accelerating Long Context Length LLM Inference](https://mlsanity.com/paper/2411.09688/) (2024)
- [LongMamba: Enhancing Mamba's Long Context Capabilities via Training-Free Receptive Field Enlargement](https://mlsanity.com/paper/2504.16053/) (2025)
- [D2O: Dynamic Discriminative Operations for Efficient Long-Context Inference of Large Language Models](https://mlsanity.com/paper/2406.13035/) (2024)
- [Cognitive Memory in Large Language Models](https://mlsanity.com/paper/2504.02441/) (2025)
- [QuantSpec: Self-Speculative Decoding with Hierarchical Quantized KV Cache](https://mlsanity.com/paper/2502.10424/) (2025)
- [Locret: Enhancing Eviction in Long-Context LLM Inference with Trained Retaining Heads on Consumer-Grade Devices](https://mlsanity.com/paper/2410.01805/) (2024)
- [Homogeneous Keys, Heterogeneous Values: Exploiting Local KV Cache Asymmetry for Long-Context LLMs](https://mlsanity.com/paper/2506.05410/) (2025)
- [KV-Latent: Dimensional-level KV Cache Reduction with Frequency-aware Rotary Positional Embedding](https://mlsanity.com/paper/2507.11273/) (2025)
- [An Evolved Universal Transformer Memory](https://mlsanity.com/paper/2410.13166/) (2024)
- [NOSA: Native and Offloadable Sparse Attention](https://mlsanity.com/paper/2510.13602/) (2025)
- [KV Admission: Learning What to Write for Efficient Long-Context LLM Inference](https://mlsanity.com/paper/2512.17452/) (2025)
- [Directed Information γ-covering: An Information-Theoretic Framework for Context Engineering](https://mlsanity.com/paper/2510.00079/) (2025)
- [Towards Sampling Data Structures for Tensor Products in Turnstile Streams](https://mlsanity.com/paper/2510.03678/) (2025)
- [FlowNar: Scalable Streaming Narration for Long-Form Videos](https://mlsanity.com/paper/2606.00620/) (2026)
- [Reference Trustable Decoding: A Training-Free Augmentation Paradigm for Large Language Models](https://mlsanity.com/paper/2409.20181/) (2024)
- [Token Entropy Regularization for Multi-modal Antenna Affiliation Identification](https://mlsanity.com/paper/2601.21280/) (2026)

## Abstract (arXiv)

As Large Language Models (LLMs) become increasingly prevalent in various domains, their ability to process inputs of any length and maintain a degree of memory becomes essential. However, the one-off input of overly long texts is limited, as studies have shown that when input lengths exceed the LLMs' pre-trained text length, there is a dramatic decline in text generation capabilities. Moreover, simply extending the length of pre-training texts is impractical due to the difficulty in obtaining long text data and the substantial memory consumption costs this would entail for LLMs. Recent efforts have employed streaming inputs to alleviate the pressure of excessively long text inputs, but this approach can significantly impair the model's long-term memory capabilities. Motivated by this challenge, we introduce Streaming Infinite Retentive LLM (SirLLM), which allows LLMs to maintain longer memory during infinite-length dialogues without the need for fine-tuning. SirLLM utilizes the Token Entropy metric and a memory decay mechanism to filter key phrases, endowing LLMs with both long-lasting and flexible memory. We designed three distinct tasks and constructed three datasets to measure the effectiveness of SirLLM from various angles: (1) DailyDialog; (2) Grocery Shopping; (3) Rock-Paper-Scissors. Our experimental results robustly demonstrate that SirLLM can achieve stable and significant improvements across different LLMs and tasks, compellingly proving its effectiveness. When having a coversation, "A sir could forget himself," but SirLLM never does! Our code is publicly available at https://github.com/Zoeyyao27/SirLLM