Explore our ablation studies using text-bison and PaLM 2-L to uncover critical design choices in prompt optimization.

Explore the results of prompt optimization on GSM8K and BBH tasks.

Explore the impact of OPRO on LLM performance across GSM8K and BBH benchmarks.

Discover how meta-prompt design enhances prompt optimization for LLMs in tasks like GSM8K.

Explore how OPRO enhances prompt optimization for natural language tasks by maximizing accuracy through innovative setups.

Explore how LLMs perform when faced with the Traveling Salesman Problem (TSP).

How do LLMs perform as optimizers in mathematical optimization?

Meta-prompts guide LLMs in the OPRO framework, detailing optimization problems and leveraging past solutions.

OPRO leverages large language models (LLMs) to optimize tasks through iterative steps.

Explore how Optimization by PROmpting uses LLMs as derivative-free optimizers reaching up to 50% accuracy improvements over human prompts in complex tasks.

Explore the meta-prompt designed for math optimization, outlining its structure and effectiveness in guiding LLMs to generate better solutions for math problems

Learn about common failures encountered with large language models (LLMs) during optimization tasks.

Explore prompting formats for scorer LLMs, highlighting examples of Q_begin, Q_end, and A_begin formats in relation to the "QA" pattern.

This paper highlights the use of large language models (LLMs) as optimizers in prompt optimization.

Review tabulated instructions for prompt optimization on BBH tasks comparing results from PaLM 2-L-IT and GPT-3.5-turbo optimizers against established baselines

Explore the upward trends observed in prompt optimization curves across 21 BBH tasks using the text-bison scorer and PaLM 2-L-IT optimizer.

Explore the tailored meta-prompts for different optimizer models, including PaLM 2-L and GPT models, and their effectiveness in prompt optimization tasks.

Pros, cons, and comparisons of various large language models (LLM) for informed decision-making in selecting models for evaluation.

This post explores AI accelerators and their impact on deploying Large Language Models (LLMs) at scale.

Exploration of model architecture optimizations for Large Language Model (LLM) inference, focusing on Group Query Attention (GQA) and Mixture of Experts (MoE)