Published April 22, 2026
| Version FSE-2026-camera-ready
Software
Open
Artifact for "Clotho: Measuring Task-Specific Pre-Generation Test Adequacy for LLM Inputs"
Authors/Creators
- 1. Korea Advanced Institute of Science and Technology
- 2. Chalmers University of Technology
Description
This artifact accompanies the paper "Clotho: Measuring Task-Specific Pre-Generation Test Adequacy for LLM Inputs" (FSE 2026).
The clotho-main.zip archive includes:
- Scripts for hidden state extraction
- Prompt templates and test suites used in the experiments, along with pipelines for synthetic data construction
- Implementation of the proposed method (Clotho) and baseline pre/post-generation test adequacy metrics
- Installation and execution guide provided as
README.md
The data.tar.gz archive includes:
- Generated LLM inference outputs across different open-source models and proprietary models
- Score outputs for baseline methods
The results_clotho_10seeds_fse2026.tar.gz archive includes:
- Intermediate outputs collected at each iteration of Clotho