Prompt tuning, ICL and friends // best practices
LLMs are good at in-context learning // more specific, precise and parameterized prompts (examples)+ RAG yield better results.
Each prompt has specific domain needs and parameters that must be fine-tuned for better retrieval augmented generation (RAG).
GraphRAG example:
Prompt Tuning ⚙️
GraphRAG provides the ability to create domain adapted prompts for the generation of the knowledge graph. This step is optional, though it is highly encouraged to run it as it will yield better results when executing an Index Run.
These are generated by loading the inputs, splitting them into chunks (text units) and then running a series of LLM invocations and template substitutions to generate the final prompts. We suggest using the default values provided by the script, but in this page you’ll find the detail of each in case you want to further explore and tweak the prompt tuning algorithm.
Retrieval-augmented generation (RAG) techniques have proven to be effective in integrating up-to-date information, mitigating hallucinations, and enhancing response quality, particularly in specialized domains. While many RAG approaches have been proposed to enhance large language models through query-dependent retrievals, these approaches still suffer from their complex implementation and prolonged response times. Typically, a RAG workflow involves multiple processing steps, each of which can be executed in various ways. Here, we investigate existing RAG approaches and their potential combinations to identify optimal RAG practices. Through extensive experiments, we suggest several strategies for deploying RAG that balance both performance and efficiency. Moreover, we demonstrate that multimodal retrieval techniques can significantly enhance question-answering capabilities about visual inputs and accelerate the generation of multimodal content using a “retrieval as generation” strategy. Resources are available at https://github.com/FudanDNN-NLP/RAG.
Large language models (LLMs) face significant challenges stemming from their inherent limitations in knowledge, memory, alignment, and action. These challenges cannot be addressed by LLMs alone, but should rely on assistance from the external world, such as knowledge base, memory store, demonstration examples, and tools. Retrieval augmentation stands as a vital mechanism for bridging the gap between LLMs and the external assistance. However, conventional methods encounter two pressing issues. On the one hand, the general-purpose retrievers are not properly optimized for the retrieval augmentation of LLMs. On the other hand, the task-specific retrievers lack the required versatility, hindering their performance across the diverse retrieval augmentation scenarios. In this work, we present a novel approach, the LLM-Embedder, which comprehensively supports the diverse retrieval augmentation needs of LLMs with one unified embedding model. Training such a unified model is non-trivial, as various retrieval tasks aim to capture distinct semantic relationships, often subject to mutual interference. To address this challenge, we systematically optimize our training methodology. This includes reward formulation based on LLMs’ feedback, the stabilization of knowledge distillation, multi-task fine-tuning with explicit instructions, and homogeneous in-batch negative sampling. These optimization strategies contribute to the outstanding empirical performance of the LLM-Embedder. Notably, it yields remarkable enhancements in retrieval augmentation for LLMs, surpassing both general-purpose and task-specific retrievers in various evaluation scenarios. Our checkpoint and source code are publicly available at https://github.com/FlagOpen/FlagEmbedding
