NSF National Deep Inference Fabric Platform Access for LLM Transparency Research

NSF National Deep Inference Fabric Platform Access for LLM Transparency Research is sponsored by National Science Foundation (NSF). The National Deep Inference Fabric (NDIF) is an NSF-funded ($9 million) collaborative research platform at Northeastern University providing US researchers free remote access to large pretrained AI models for detailed and reproducible experiments on model internals.

NSF | National Deep Inference Fabric NDIF is now available — free remote access to large-scale AI models for research. Get started → Workbench UI System Status The NSF National Deep Inference Fabric provides free remote access to large-scale AI models, enabling researchers and students to perform transparent, reproducible experiments on model internals. Data sourced from NDIF internal metrics and public GitHub repository statistics.

The National Deep Inference Fabric is a unique nationwide research computing fabric that enables scientists to perform transparent and reproducible experiments on the largest-scale open AI systems. NDIF has three parts: A nationwide high-performance computing fabric powered by NCSA's Delta — utilizing one 8xH200 node and six 4xA40 nodes — providing free remote access to run experiments on large-scale AI models.

An open-source PyTorch-based toolkit (850+ GitHub stars) that lets researchers inspect, modify, and customize internal computations of AI models, complete with remote access to large scale models. A nationwide training program developed with PIT-UN, a consortium of 63 universities and colleges, providing workshops, tutorials, and resources to build a broad community of AI researchers.

Learn more about the Fabric → Access and inspect open-source model internals remotely on NDIF with the NNsight API. Deploy models on-demand with full transparency into their computations. All models are free for research use — no GPU required on your end.

View All Models NNsight Docs ↗ meta-llama/Meta-Llama-3. 1-70B meta-llama/Meta-Llama-3. 1-8B meta-llama/Meta-Llama-3.

1-405B Researchers worldwide use NDIF and NNsight to uncover how large-scale AI models work, with 110 + published papers at top venues including ICLR, NeurIPS, ICML, and EMNLP.

Activation Steering via Generative Causal Mediation Aruna Sankaranarayanan, Amir Zur, Atticus Geiger, Dylan Hadfield-Menell Researchers tackle the challenge of pinpointing and controlling specific behaviors in language models, which can be dispersed throughout lengthy responses.

By developing a new approach, they aim to provide more precise control over these models, enabling more effective intervention and modification of their outputs. This work has implications for improving the reliability and trustworthiness of language models in real-world applications. Can you map it to English?

The Role of Cross-Lingual Alignment in the Multilingual Performance of LLMs Kartik Ravisankar, HyoJung Han, Sarah Wiegreffe, Marine Carpuat Researchers investigate the role of linguistic and cultural biases in machine translation systems, shedding light on how these biases can affect the accuracy and fairness of translations.

By examining the intersection of language, culture, and technology, this study aims to promote more inclusive and equitable language technologies. Its findings have implications for the development of more culturally sensitive machine translation systems. Counting Hypothesis: Potential Mechanism of In-Context Learning Jung H.

Lee, Sujith Vijayan Researchers are working to understand how large language models can learn specific tasks from just a few examples, a phenomenon known as In-Context Learning. This ability has the potential to expand the use of these models into new areas, but its underlying mechanisms are still not well understood, making it difficult to correct errors or diagnose issues.

By proposing a new hypothesis, the "counting hypothesis," this study aims to shed light on how these models support In-Context Learning.

DFWe: Efficient Knowledge Distillation of Fine-tuned Whisper Encoder for Speech Emotion Recognition Y Ma, X Jiang, J Sang, R Li Researchers are exploring ways to adapt powerful pre-trained speech models, like Whisper, to recognize emotions in speech, a task that requires more nuanced understanding than just acoustic modeling.

By addressing the limitations of these models in capturing emotional cues, this work aims to improve speech emotion recognition. This advancement could lead to more empathetic and human-like interactions with voice-based systems.

Disentangling meaning from language in LLM-based machine translation Théo Lasnier, Armel Zebaze, Djamé Seddah, Rachel Bawden, Benoît Sagot Researchers are working to understand how Large Language Models (LLMs) work by developing new methods for Mechanistic Interpretability (MI), which aims to explain the inner workings of neural networks. This effort is crucial for building trust in AI systems and identifying potential biases or flaws.

By shedding light on how LLMs process and generate language, MI can help improve the reliability and transparency of these powerful models. Do Transformers Use their Depth Adaptively? Evidence from a Relational Reasoning Task Alicia Curth, Rachel Lawrence, Sushrut Karmalkar, Niranjani Prasad Researchers explore how transformers, a type of AI model, adapt their processing depth to tackle tasks of varying complexity.

By analyzing how these models process information across different layers and tokens, they shed light on the strategies transformers use to solve problems. The study reveals that transformers can adapt their depth to suit the task at hand, especially when fine-tuned for specific tasks, allowing them to efficiently allocate processing resources.

Open-source repositories using NDIF and NNsight — production-grade interpretability tools, libraries, and research code from the community. callummcdougall / ARENA_3. 0 Jupyter Notebook 1,124 728 hijohnnylin / neuronpedia open source interpretability platform 🧠 saprmarks / dictionary_learning Active 10 months ago · MIT Delphi was the home of a temple to Phoebus Apollo, which famously had the inscription, 'Know Thyself.'

This library lets language models know themselves through automated interpretability. Active 4 days ago · Apache-2. 0 TransluceAI / observatory A toolkit for describing model features and intervening on those features to steer behavior.

Active 3 months ago · MIT saprmarks / feature-circuits Active 8 months ago · MIT Three simple steps to start running experiments on large-scale AI models. Install the open-source NNsight library with a single pip command. Create a free account to get remote access to large-scale models hosted on NDIF.

Run transparent, reproducible experiments on model internals — no GPU required. $ model. trace(remote=True) We'd love to have you.

Whether you're debugging your first experiment or contributing to the codebase, there's a place for you here. Got a question? Stuck on an experiment?

Our Discord is where researchers share ideas, get unstuck, and geek out about model internals. Learn by doing. We run regular workshops that take you from zero to running your first remote experiment on large-scale models.

NNsight is open source and built by researchers like you. Whether it's code, documentation, or a bug report—every contribution helps.