Prompt Codebooks：面向语言模型指令优化的离散组合式优化方法

View PDF HTML (experimental)

Abstract:Automatic prompt optimization (APO) has driven significant gains in LLM-based agentic workflows. However, most existing methods treat each task's prompt as a monolithic, instance-blind string optimized through global edits, producing brittle updates and preventing the reuse of learned sub-behaviors. We propose Prompt Codebook Optimization (PCO), a novel compositional prompt optimization framework that recasts APO as discrete learning over a finite vocabulary of natural-language instincts--atomic, reusable instruction units. PCO organizes prompt-construction knowledge in a discrete codebook and routes each input to a small subset of entries via an LLM-based encoder; a generator composes them into a prompt for the executor; a critic emits a structured verdict that decomposes by attribution into per-variable textual gradients, jointly training the encoder, generator, critic, and codebook under a language-valued min-max objective. The resulting routing is per-instance: different inputs in the same task receive different instinct compositions. Across six benchmarks, PCO improves aggregate performance over zero-shot by +13.50 points on Qwen3-8B and +11.80 points on LLaMA-3.1-8B. Crucially, PCO surpasses GEPA on HotpotQA by +6.34 points (Qwen3-8B) and +4.27 points (LLaMA-3.1-8B), while simultaneously reducing deployed prompt length by up to 14.1$\times$ vs. MIPROv2 and 3.0$\times$ vs. GEPA.

Subjects:	Artificial Intelligence (cs.AI)
Cite as:	arXiv:2605.28360 [cs.AI]
	(or arXiv:2605.28360v3 [cs.AI] for this version)
	https://doi.org/10.48550/arXiv.2605.28360 arXiv-issued DOI via DataCite

Submission history

From: Jyotirmoy Nath [view email]
[v1] Wed, 27 May 2026 11:57:12 UTC (1,179 KB)
[v2] Mon, 3 Aug 2026 14:49:01 UTC (113 KB)
[v3] Fri, 16 Oct 2026 16:00:25 UTC (994 KB)