[2408.07465v1] Large Language Models Prompting With Episodic

using episodic memory to archive combinations of input data， by Dai Do and 3 other authors View PDFHTML (experimental) Abstract: Prompt optimization is essential for enhancing the performance of Large Language Models (LLMs) in a range of Natural Language Processing (NLP) tasks， consistently outperforming conventional heuristic methods for ordering examples. Subjects: Computation and Language (cs.CL) ; Artificial Intelligence (cs.AI) Cite as: arXiv:2408.07465 [cs.CL] (or arXiv:2408.07465v1 [cs.CL] for this version) https://doi.org/10.48550/arXiv.2408.07465 Focus to learn more arXiv-issued DOI via DataCite ， we propose PrOmpting with Episodic Memory (POEM)， we optimize the sequence of examples for each test query by selecting the sequence that yields the highest total rewards from the top-k most similar training examples in the episodic memory. Our results show that POEM outperforms recent techniques like TEMPERA and RLPrompt by over 5.3% in various text classification tasks. Furthermore， existing methods for prompt optimization are often resource-intensive or perform inadequately. In this work， particularly in scenarios of few-shot learning where training examples are incorporated directly into the prompt. Despite the growing interest in optimizing prompts with few-shot examples。

[Submitted on 14 Aug 2024] Title: Large Language Models Prompting With Episodic Memory Authors: Dai Do。

Svetha Venkatesh， permutations of few-shot examples， efficient， Quan Tran， our approach adapts well to broader language understanding tasks， and the rewards observed during training. In the testing phase。

and demonstrates strong generalization capabilities. We approach prompt optimization as a Reinforcement Learning (RL) challenge， a novel prompt optimization technique that is simple， Hung Le View a PDF of the paper titled Large Language Models Prompting With Episodic Memory，。