4 papers
Reward Hacking as Equilibrium under Finite Evaluation
Jiacheng Wang, Jinbin Huang
We prove that under five minimal axioms -- multi-dimensional quality, finite evaluation, effective optimization, resource finiteness, and combinatorial interaction -- any optimized…
PromptAid: Prompt Exploration, Perturbation, Testing and Iteration using Visual Analytics for Large Language Models
Aditi Mishra, Utkarsh Soni, Anjana Arunkumar +3
Large Language Models (LLMs) have gained widespread popularity due to their ability to perform ad-hoc Natural Language Processing (NLP) tasks with a simple natural language prompt.…
InFiConD: Interactive No-code Fine-tuning with Concept-based Knowledge Distillation
Jinbin Huang, Wenbin He, Liang Gou +2
The emergence of large-scale pre-trained models has heightened their application in various downstream tasks, yet deployment is a challenge in environments with limited computation…
InterVLS: Interactive Model Understanding and Improvement with Vision-Language Surrogates
Jinbin Huang, Wenbin He, Liang Gou +2
Deep learning models are widely used in critical applications, highlighting the need for pre-deployment model understanding and improvement. Visual concept-based methods, while inc…