# FlashBack: Efficient Retrieval-Augmented Language Modeling for Fast Inference

arXiv: https://arxiv.org/abs/2405.04065 · HTML: https://mlsanity.com/paper/2405.04065/

## Key facts

- Cited by 1 papers in the map (data through 14 Oct 2026).
- Placed in Generation and context integration, one of 15 directions in Retrieval-augmented generation.
- Builds on 22 of its 31 references that are in the map.
- First posted on arXiv 7 May 2024.

## Placed in

- Retrieval-augmented generation / [Generation and context integration](https://mlsanity.com/rag/generation-and-context-integration/)

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

- [Language Models are Few-Shot Learners](https://mlsanity.com/paper/2005.14165/) (2020)
- [LoRA: Low-Rank Adaptation of Large Language Models](https://mlsanity.com/paper/2106.09685/) (2021)
- [Retrieval-Augmented Generation for Knowledge-Intensive NLP Tasks](https://mlsanity.com/paper/2005.11401/) (2020)
- [Dense Passage Retrieval for Open-Domain Question Answering](https://mlsanity.com/paper/2004.04906/) (2020)
- [OPT: Open Pre-trained Transformer Language Models](https://mlsanity.com/paper/2205.01068/) (2022)
- [Pointer Sentinel Mixture Models](https://mlsanity.com/paper/1609.07843/) (2016)
- [TriviaQA: A Large Scale Distantly Supervised Challenge Dataset for Reading Comprehension](https://mlsanity.com/paper/1705.03551/) (2017)
- [Locating and Editing Factual Associations in GPT](https://mlsanity.com/paper/2202.05262/) (2022)
- [The Pile: An 800GB Dataset of Diverse Text for Language Modeling](https://mlsanity.com/paper/2101.00027/) (2020)
- [REALM: Retrieval-Augmented Language Model Pre-Training](https://mlsanity.com/paper/2002.08909/) (2020)
- [Improving language models by retrieving from trillions of tokens](https://mlsanity.com/paper/2112.04426/) (2021)
- [REPLUG: Retrieval-Augmented Black-Box Language Models](https://mlsanity.com/paper/2301.12652/) (2023)
- [FinanceBench: A New Benchmark for Financial Question Answering](https://mlsanity.com/paper/2311.11944/) (2023)
- [Generative Representational Instruction Tuning](https://mlsanity.com/paper/2402.09906/) (2024)
- [RA-DIT: Retrieval-Augmented Dual Instruction Tuning](https://mlsanity.com/paper/2310.01352/) (2023)
- [UPRISE: Universal Prompt Retrieval for Improving Zero-Shot Evaluation](https://mlsanity.com/paper/2303.08518/) (2023)
- [Reliable, Adaptable, and Attributable Language Models with Retrieval](https://mlsanity.com/paper/2403.03187/) (2024)
- [Augmentation-Adapted Retriever Improves Generalization of Language Models as Generic Plug-In](https://mlsanity.com/paper/2305.17331/) (2023)
- [Shall We Pretrain Autoregressive Language Models with Retrieval? A Comprehensive Study](https://mlsanity.com/paper/2304.06762/) (2023)
- [InstructRetro: Instruction Tuning post Retrieval-Augmented Pretraining](https://mlsanity.com/paper/2310.07713/) (2023)
- [RAVEN: In-Context Learning with Retrieval-Augmented Encoder-Decoder Language Models](https://mlsanity.com/paper/2308.07922/) (2023)
- [Context Compression for Auto-regressive Transformers with Sentinel Tokens](https://mlsanity.com/paper/2310.08152/) (2023)

## Built on by (most cited first, 1 of 1)

- [CogGuard: Cognitive and Operational Profiling for Proactive Warning in Edge Intelligent Services](https://mlsanity.com/paper/2606.15199/) (2026)

## Abstract (arXiv)

Retrieval-Augmented Language Modeling (RALM) by integrating large language models (LLM) with relevant documents from an external corpus is a proven method for enabling the LLM to generate information beyond the scope of its pre-training corpus. Previous work utilizing retrieved content by simply prepending it to the input poses a high runtime issue, which degrades the inference efficiency of the LLMs because they fail to use the Key-Value (KV) cache efficiently. In this paper, we propose FlashBack, a modular RALM designed to improve the inference efficiency of RALM with appending context pattern while maintaining decent performance after fine-tuning by Low-Rank Adaption. FlashBack appends retrieved documents at the end of the context for efficiently utilizing the KV cache instead of prepending them. And we introduce Marking Token as two special prompt tokens for marking the boundary of the appending context during fine-tuning. Our experiments on testing generation quality show that FlashBack can remain decent generation quality in perplexity. And the inference speed of FlashBack is up to 4× faster than the prepending counterpart on a 7B LLM (Llama 2) in the runtime test. Via bypassing unnecessary re-computation, it demonstrates an advancement by achieving significantly faster inference speed, and this heightened efficiency will substantially reduce inferential cost.