metadata
base_model:
- meta-llama/Llama-3.1-8B-Instruct
datasets:
- infosense/yield
language:
- en
license: llama3.1
library_name: peft
pipeline_tag: text-generation
YIELD Fine-Tuning Adapters
This repository contains the persona adapter models presented in the paper YIELD: A Large-Scale Dataset and Evaluation Framework for Information Elicitation Agents.
Model Information
All adapters in this repository are LoRA adapters trained on top of the Llama-3.1-8B-Instruct, Llama-3.2-3B-Instruct, and DeepSeek-R1-Distill-Llama-8B models. All models are fine-tuned using both the Supervised Fine-Tuning (SFT) and Offline Reinforcement-Learning (ORL) pipelines detailed in the paper. These models are designed to act as Information Elicitation Agents (IEAs), which aim to elicit information from users to support institutional or task-oriented objectives.
Resources
- Code Repository: GitHub - infosenselab/yield
- Dataset: infosense/yield
- Paper: Hugging Face Papers
Citing YIELD
If you use this resource in your projects, please cite the following paper:
@misc{De_Lima_YIELD_A_Large-Scale_2026,
author = {De Lima, Victor and Yang, Grace Hui},
doi = {10.48550/arXiv.2604.10968},
title = {{YIELD: A Large-Scale Dataset and Evaluation Framework for Information Elicitation Agents}},
url = {https://arxiv.org/abs/2604.10968},
year = {2026}
}