mlsanityData to 14 Oct 2026

eDIF: A European Deep Inference Fabric for Remote Interpretability of LLM

arXiv 2508.10553 ↗First posted 14 Aug 2025cs.CLcited by 1

On the map: Mechanistic interpretability → Knowledge and representations inside LLMs

Abstract (arXiv)

This paper presents a feasibility study on the deployment of a European Deep Inference Fabric (eDIF), an NDIF-compatible infrastructure designed to support mechanistic interpretability research on large language models. The need for widespread accessibility of LLM interpretability infrastructure in Europe drives this initiative to democratize advanced model analysis capabilities for the research community. The project introduces a GPU-based cluster hosted at Ansbach University of Applied Sciences and interconnected with partner institutions, enabling remote model inspection via the NNsight API. A structured pilot study involving 16 researchers from across Europe evaluated the platform's technical performance, usability, and scientific utility. Users conducted interventions such as activation patching, causal tracing, and representation analysis on models including GPT-2 and DeepSeek-R1-70B. The study revealed a gradual increase in user engagement, stable platform performance throughout, and a positive reception of the remote experimentation capabilities. It also marked the starting point for building a user community around the platform. Identified limitations such as prolonged download durations for activation data as well as intermittent execution interruptions are addressed in the roadmap for future development. This initiative marks a significant step towards widespread accessibility of LLM interpretability infrastructure in Europe and lays the groundwork for broader deployment, expanded tooling, and sustained community collaboration in mechanistic interpretability research.

Key facts. Cited by 1 paper in the map (data through 14 Oct 2026). Placed in Knowledge and representations inside LLMs, one of 13 directions in Mechanistic interpretability. Builds on 9 of its 32 references that are in the map. First posted on arXiv 14 Aug 2025.

Builds on

9 of 32 references.

References · 9

  1. 2024Sparse Feature Circuits: Discovering and Editing Interpretable Causal Graphs in Language Modelscited by 435
  2. 2024Mechanistic Interpretability for AI Safety -- A Reviewcited by 422
  3. 2024How to use and interpret activation patchingcited by 214
  4. 2024LLM Circuit Analyses Are Consistent Across Training and Scalecited by 63
  5. 2024pyvene: A Library for Understanding and Improving PyTorch Models via Interventionscited by 54
  6. 2024NNsight and NDIF: Democratizing Access to Open-Weight Foundation Model Internalscited by 45
  7. 2024The 2024 Foundation Model Transparency Indexcited by 19
  8. 2025An explainable transformer circuit for compositional generalizationcited by 8
  9. 2025Comprehensive Analysis of Transparency and Accessibility of ChatGPT, DeepSeek, And other SoTA Large Language Modelscited by 8

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2026: 1.

1 paper.

  1. 2026How Open Must Language Models be to Enable Reliable Scientific Inference?cited by 2

Similar, not cited

Close in wording, no citation either way

  1. 2025Challenges and Applications of Large Language Models: A Comparison of GPT and DeepSeek family of models
  2. 2022Augmenting Interpretable Models with LLMs during Training
  3. 2024Rethinking Interpretability in the Era of Large Language Models
  4. 2025Adapting Language-Specific LLMs to a Reasoning Model in One Day via Model Merging -- An Open Recipe
  5. 2024DeepSeek LLM: Scaling Open-Source Language Models with Longtermism
  6. 2024LLM Inference Unveiled: Survey and Roofline Model Insights
  7. 2023DeepDecipher: Accessing and Investigating Neuron Activation in Large Language Models
  8. 2024Highly Optimized Kernels and Fine-Grained Codebooks for LLM Inference on Arm CPUs