Research Tools: code, datasets, & models

interwhen is a test-time verification framework for language models that enforces correctness with respect to a set of verifiers. It is designed to improve instance-level reliability in reasoning systems, particularly in high-stakes domains where occasional…

Declarative, interactive data documents Chartifact is a low-code document format for creating interactive, data-driven pages such as reports, dashboards, and presentations. It travels like a document and works like a mini app. Designed for use…

A security-focused library OS [!NOTE]This project is currently actively evolving and improving. While we areworking toward a stable release, some APIs and interfaces may change as thedesign continues to mature. You are welcome to explore…

This repository is the official implementation of the paper “TestExplora: Benchmarking LLMs for Proactive Bug Discovery via Repository-Level Test Generation” It can be used for baseline evaluation using the prompts mentioned in the paper. TestExplora…

Scaling-Aware Best-of-N Estimation of Risk A Python package for predicting large-scale adversarial risk in Large Language Models under Best-of-N sampling. Paper: https://arxiv.org/pdf/2601.22636 (opens in new tab)   Standard LLM safety evaluations use single-shot (ASR@1) metrics,…

SigmaCollab is a dataset that enables research on human-AI physically situated collaboration. The dataset consists of a set of 85 sessions in which untrained participants were guided by a mixed-reality assistive AI agent in performing…

Generative modeling, representation learning, and classification are three core problems in machine learning (ML), yet their state-of-the-art (SoTA) solutions remain largely disjoint. In this paper, we ask: Can a unified principle address all three? Such…

We propose Large Market Model (LMM), an order-level generative foundation model, for financial market simulation, akin to language modeling in the digital world. Our financial Market Simulation engine (MarS), powered by LMM, addresses the domain-specific…

MIRA is a foundation model for medical time-series, designed to learn a unified representation space across heterogeneous clinical datasets and support zero-shot forecasting in real-world healthcare settings. Unlike conventional time-series models that operate on fixed…

OptiMind-SFT is a specialized 20B parameter model designed to bridge the gap between natural language and executable optimization solvers. It automates the translation of complex decision-making problems—such as supply chain planning, scheduling, and resource allocation—into…