6 papers
Hybrid Neural World Models
Pranav Lakshmanan, Paras Chopra
Neural surrogates promise large speedups over classical solvers for physical dynamics but fail silently at sharp dynamical events such as shocks, fronts, and contact. We present hy…
Discovering Reinforcement Learning Interfaces with Large Language Models
Akshat Singh Jaswal, Ashish Baghel, Paras Chopra
Reinforcement learning systems rely on environment interfaces that specify observations and reward functions, yet constructing these interfaces for new tasks often requires substan…
See, Symbolize, Act: Grounding VLMs with Spatial Representations for Better Gameplay
Ashish Baghel, Paras Chopra
Vision-Language Models (VLMs) excel at describing visual scenes, yet struggle to translate perception into precise, grounded actions. We investigate whether providing VLMs with bot…
Building Interpretable Models for Moral Decision-Making
Mayank Goel, Aritra Das, Paras Chopra
We build a custom transformer model to study how neural networks make moral decisions on trolley-style dilemmas. The model processes structured scenarios using embeddings that enco…
IPO: Your Language Model is Secretly a Preference Classifier
Shivank Garg, Ayush Singh, Shweta Singh +1
Reinforcement learning from human feedback (RLHF) has emerged as the primary method for aligning large language models (LLMs) with human preferences. While it enables LLMs to achie…
Do GFlowNets Transfer? Case Study on the Game of 24/42
Adesh Gupta, Abhinav Kumar, Mansi Gupta +1
Generating diverse solutions is key to human-like reasoning, yet autoregressive language models focus on single accurate responses, limiting creativity. GFlowNets optimize solution…