# Auto-Prompt Generation is Not Robust: Prompt Optimization Driven by Pseudo Gradient

arXiv: https://arxiv.org/abs/2412.18196 · HTML: https://mlsanity.com/paper/2412.18196/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 24 Dec 2024.

## Built on by (most cited first, 6 of 6)

- [A Comprehensive Survey in LLM(-Agent) Full Stack Safety: Data, Training and Deployment](https://mlsanity.com/paper/2504.15585/) (2025)
- [Meaningless Tokens, Meaningful Gains: How Activation Shifts Enhance LLM Reasoning](https://mlsanity.com/paper/2510.01032/) (2025)
- [PQR: A Framework to Generate Diverse and Realistic User Queries that Elicit QA Agent Failures](https://mlsanity.com/paper/2605.16551/) (2026)
- [Towards Spec Learning: Inference-Time Alignment from Preference Pairs](https://mlsanity.com/paper/2606.24004/) (2026)
- [Prompt-Robust Language Models: Which Training Strategies Work?](https://mlsanity.com/paper/2609.01217/) (2026)
- [Do Prompt Optimizers Break Requirements They Are Not Scored On? A Preregistered Study](https://mlsanity.com/paper/2610.13287/) (2026)

## Abstract (arXiv)

While automatic prompt generation methods have recently received significant attention, their robustness remains poorly understood. In this paper, we introduce PertBench, a comprehensive benchmark dataset that includes a wide range of input perturbations, designed to systematically evaluate the robustness of current auto-prompting techniques. Our analysis reveals substantial vulnerabilities in existing prompt generation strategies, where even minor modifications to the prompt can lead to significant differences in model output. To address this issue, we propose PGO, a gradient-free prompt generation framework that leverages perturbation types as pseudo-gradient signals to guide LLMs in producing more robust prompts. In contrast to existing methods that assess prompt quality only on clean, well-structured inputs, our approach explicitly emphasizes robustness under noisy and perturbed conditions. Extensive experiments across diverse tasks and multiple LLMs show PGO consistently outperforms previous methods in maintaining performance under input perturbations.