CELI: Controller-Embedded Language Model Interactions

We introduce Controller-Embedded Language Model Interactions (CELI), a framework that integrates control logic directly within language model (LM) prompts, facilitating complex, multi-stage task execution. CELI addresses limitations of existing prompt engineering and workflow optimization techniques...

Full description

Saved in:
Bibliographic Details
Published in:arXiv.org 2024-10
Main Authors: Jan-Samuel, Wagner, DeCaprio, Dave, Abishek Chiffon Muthu Raja, Holman, Jonathan M, Brady, Lauren K, Cheung, Sky C, Barzekar, Hosein, Yang, Eric, Martinez, Mark Anthony, Soong, David, Sridhar, Sriram, Han, Si, Higgs, Brandon W, Hamadeh, Hisham, Ogden, Scott
Format: Article
Language:English
Subjects:
Adaptive control
Benchmarks
Computer programming
Controllers
Domains
Embedding
Encyclopedias
Optimization
Optimization techniques
Prompt engineering
Workflow
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:We introduce Controller-Embedded Language Model Interactions (CELI), a framework that integrates control logic directly within language model (LM) prompts, facilitating complex, multi-stage task execution. CELI addresses limitations of existing prompt engineering and workflow optimization techniques by embedding control logic directly within the operational context of language models, enabling dynamic adaptation to evolving task requirements. Our framework transfers control from the traditional programming execution environment to the LMs, allowing them to autonomously manage computational workflows while maintaining seamless interaction with external systems and functions. CELI supports arbitrary function calls with variable arguments, bridging the gap between LMs' adaptive reasoning capabilities and conventional software paradigms' structured control mechanisms. To evaluate CELI's versatility and effectiveness, we conducted case studies in two distinct domains: code generation (HumanEval benchmark) and multi-stage content generation (Wikipedia-style articles). The results demonstrate notable performance improvements across a range of domains. CELI achieved a 4.9 percentage point improvement over the best reported score of the baseline GPT-4 model on the HumanEval code generation benchmark. In multi-stage content generation, 94.4% of CELI-produced Wikipedia-style articles met or exceeded first draft quality when optimally configured, with 44.4% achieving high quality. These outcomes underscore CELI's potential for optimizing AI-driven workflows across diverse computational domains.
ISSN:2331-8422