Guardado en:
Detalles Bibliográficos
Autores principales: George, Sonny, Sypherd, Chris, Cashman, Dylan
Formato: Preprint
Publicado: 2024
Materias:
Acceso en línea:https://arxiv.org/abs/2411.12828
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866909396387233792
author George, Sonny
Sypherd, Chris
Cashman, Dylan
author_facet George, Sonny
Sypherd, Chris
Cashman, Dylan
contents Large language model (LLM) agents show promise in an increasing number of domains. In many proposed applications, it is expected that the agent reasons over accumulated experience presented in an input prompt. We propose the OEDD (Operationalize Experience Despite Distraction) corpus, a human-annotator-validated body of scenarios with pre-scripted agent histories where the agent must make a decision based on disparate experiential information in the presence of a distractor. We evaluate three state-of-the-art LLMs (GPT-3.5 Turbo, GPT-4o, and Gemini 1.5 Pro) using a minimal chain-of-thought prompting strategy and observe that when (1) the input context contains over 1,615 tokens of historical interactions, (2) a crucially decision-informing premise is the rightful conclusion over two disparate environment premises, and (3) a trivial, but distracting red herring fact follows, all LLMs perform worse than random choice at selecting the better of two actions. Our code and test corpus are publicly available at: https://github.com/sonnygeorge/OEDD .
format Preprint
id arxiv_https___arxiv_org_abs_2411_12828
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Probing the Capacity of Language Model Agents to Operationalize Disparate Experiential Context Despite Distraction
George, Sonny
Sypherd, Chris
Cashman, Dylan
Computation and Language
Artificial Intelligence
Large language model (LLM) agents show promise in an increasing number of domains. In many proposed applications, it is expected that the agent reasons over accumulated experience presented in an input prompt. We propose the OEDD (Operationalize Experience Despite Distraction) corpus, a human-annotator-validated body of scenarios with pre-scripted agent histories where the agent must make a decision based on disparate experiential information in the presence of a distractor. We evaluate three state-of-the-art LLMs (GPT-3.5 Turbo, GPT-4o, and Gemini 1.5 Pro) using a minimal chain-of-thought prompting strategy and observe that when (1) the input context contains over 1,615 tokens of historical interactions, (2) a crucially decision-informing premise is the rightful conclusion over two disparate environment premises, and (3) a trivial, but distracting red herring fact follows, all LLMs perform worse than random choice at selecting the better of two actions. Our code and test corpus are publicly available at: https://github.com/sonnygeorge/OEDD .
title Probing the Capacity of Language Model Agents to Operationalize Disparate Experiential Context Despite Distraction
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2411.12828