SAP Internship R&D LLM Mechanistic Interpretability Privacy - F/M 

France 

956277677

Some recent research activities focus on decomposing LLM activations into sparse linear combinations of features from a large, shared feature dictionary. Using (unsupervised) sparse autoencoder (SAE) framework allows to find such dictionaries, focusing in particular on monosemantic features (features that activate only for one lement/concept in such dictionary).• Application and replication of recent studies on mechanistic interpretability on real and synthetic datasets.
• Design of approaches to enhance privacy guarantees of individuals, by detecting sensitive personal data in various contexts (social media contents, business communications etc.).
• Design of context-dependent approaches to anonymize sensitive personal data, without compromising as much as possible on data integrity and utility.
• Create clear documentation for evaluated approaches, with the aim of producing a scientific contribution to an international venue.

• University Level: Last year of MSc in Computer Science or beyond
• Strong Python development skills
• Natural Language Processing (NLP) Knowledge
• Machine Learning Knowledge
• Fluency in English (working language)
• Abilities in organizing meeting and contacting people
• Good oral and written communication skills
• Capacity to write documents in English, ability to synthesize

: 6 months

​

Job Segment:R&D, Cloud, ERP, Intern, Research, Technology, Entry Level