Prompt Techniques

A comprehensive guide to zero-shot prompting covering instruction design, task specification, zero-shot chain-of-thought reasoning, and practical applications in large language models.

A comprehensive guide to few-shot prompting covering in-context learning, example selection strategies, demonstration formatting, and practical applications in large language models.

A comprehensive guide to automated prompt engineering covering APE techniques, meta-prompting, optimization approaches, and practical implementations for improving LLM performance.

A comprehensive guide to Chain-of-Verification prompting covering the 4-step verification process, hallucination reduction strategies, implementation methods, and real-world applications for improving LLM accuracy.

A comprehensive guide to Chain-of-Thought prompting covering foundational concepts, mathematical reasoning, implementation strategies, advanced techniques, and real-world applications in large language models.

A comprehensive guide to Active Prompting covering uncertainty-based example selection, human annotation integration, iterative improvement, and practical applications for enhancing few-shot learning performance.

A comprehensive guide to Auto-CoT prompting covering automatic demonstration construction through clustering-based diversity, zero-shot chain generation, implementation strategies, benchmark results, and integration with modern LLM workflows.

A comprehensive guide to automatic reasoning prompting techniques that enable language models to generate reasoning chains without manual demonstration design, covering Auto-CoT, ART, COSP, and related methods.

A comprehensive guide to COSP, a zero-shot prompting technique that automatically selects high-quality demonstrations from LLM's own outputs using consistency, diversity, and repetition criteria to improve reasoning without labeled data.

A comprehensive guide to Cumulative Reasoning covering the structured framework that orchestrates LLMs through Proposer, Verifier, and Reporter roles to build dynamic DAGs of verified propositions for enhanced problem-solving.

A modular approach that solves complex tasks by decomposing them into simpler sub-tasks delegated to specialized prompting-based LLMs, achieving superior performance over Chain-of-Thought and Least-to-Most prompting.

Diverse Verifier on Reasoning Steps — a comprehensive guide to enhancing LLM reasoning through diverse prompting and step-aware verification for improved accuracy on complex reasoning tasks.

A comprehensive guide to DP2O, an automated prompt optimization technique that combines dialogue-based prompt generation with policy-gradient reinforcement learning for few-shot learning tasks.

A comprehensive guide to Graph of Thoughts (GoT) prompting covering graph-based reasoning, thought transformations, aggregation operations, benchmark results, and practical applications for solving elaborate problems with large language models.