DeepTest Tool Competition 2026: Benchmarking an LLM-Based Automotive Assistant

2026-04-14

AI Summary

This work addresses the critical issue of large language model (LLM)-driven automotive manual retrieval systems frequently omitting essential safety warnings in response to user queries. To systematically evaluate this risk, the authors organize the first benchmarking competition specifically designed for LLM-based automotive assistants. The competition leverages a suite of competing tools that generate test cases through automated test generation, LLM interaction, failure detection, and diversity analysis, establishing an evaluation framework centered on failure-exposure capability and test diversity. The study assesses the effectiveness of four distinct testing tools in uncovering safety-critical flaws, offering a reproducible experimental setup, baseline results, and actionable insights into the safety assurance of LLM applications in high-stakes domains such as automotive assistance.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Safety and Robustness
• Planning, Routing, and Scheduling: Planning with Language Models

Application Category

• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information
• Economics, Online Markets and Human Computation: Cost models of using LLMs in production systems
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

This report summarizes the results of the first edition of the Large Language Model (LLM) Testing competition, held as part of the DeepTest workshop at ICSE 2026. Four tools competed in benchmarking an LLM-based car manual information retrieval application, with the objective of identifying user inputs for which the system fails to appropriately mention warnings contained in the manual. The testing solutions were evaluated based on their effectiveness in exposing failures and the diversity of the discovered failure-revealing tests. We report on the experimental methodology, the competitors, and the results.

Problem

• LLM-based automotive assistant
• failure detection
• warning omission
• test diversity
• benchmarking

Innovation

• LLM testing
• automotive assistant
• failure-revealing tests
• benchmarking
• safety warnings