AI Paper Atlas

Research map / Large language models / Agents, code and math

Learning from Guidelines: Structured Prompt Optimization for Expert Annotation Tasks

Wenliang Zhong, Haiqing Li, Thao M. Dang, Feng Jiang, Hehuan Ma, Yuzhi Guo, Jean Gao, Junzhou Huang

AAAI 2026 · Large language models › Agents, code and math · approach: LLMs for optimization and heuristics · 1 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add pipeline decomposition (ablation)GDP · NCBI-DiseaseF1: 44.6 -> 48.5Improved
add content optimization (ablation)GDP · NCBI-DiseaseF1: 42.3 -> 48.5Improved
add pipeline decomposition (ablation)GDP · CASIEF1: 80.9 -> 81.6No clear effect
add content optimization (ablation)GDP · WNUTF1: 55.2 -> 55.5No clear effect
use content-only optimization (ablation)GDP · 4 datasetsF1No clear effect
replace LLM optimizer with guideline RAS in PromptAgent (ablation)PromptAgent · HarveyNERF1: 35.0 -> 41.8Improved
replace LLM optimizer with guideline RAS in GPO (ablation)GPO · HarveyNERF1: 18.7 -> 27.3Improved
replace LLM optimizer with guideline RAS in GPO (ablation)GPO · CASIEF1: 77.4 -> 78.2Improved
use guideline-driven GDP instead of PromptAgentGPT-4 · Avg, 20-shot trainingF1: 54.4 -> 65.5Improved
use GDP instead of CoT promptingGPT-4 · Avg, 5-shot trainingF1: 58.4 -> 62.9Improved
more training examples (10 to 20 shots)PromptAgent · Avg, GPT-4F1: 56.9 -> 54.4Hurt
more training examples (10 to 20 shots)GDP · Avg, GPT-4F1: 62.5 -> 65.5Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy