Controlled Text Generation Research Internship at Naver Labs Europe

Join our research internship in ML/NLP at Naver Labs Europe on controlled text generation and contribute to innovative developments in controlled generation from LLMs.

We are offering a research internship in ML/NLP at Naver Labs Europe on controlled text generation.

The internship will focus on developing methods to impose complex constraints on the output of a large language model while preserving most of the controlled model’s distribution.

This is a key challenge in ensuring both flexibility and reliability in text generation, with applications in areas such as safety, trustworthiness or reasoning.

To tackle this, we plan to build upon the Distributional Policy Gradient method developed at Naver Labs Europe and follow up on our recent work around Guaranteed Policy (to appear at ICLR ’23).

Internship details:

• Duration: 5-6 months, starting ASAP
• Eligibility: PhD students (or outstanding Master’s students) enrolled at a university
• Location: onsite in Grenoble, France

This internship offers the opportunity to work in a cutting-edge research environment, collaborate with experienced ML/NLP researchers, and contribute to innovative developments in controlled generation from LLMs.

Learn more and apply here.

Tags: research internship, ML/NLP, controlled text generation, Distributional Policy Gradient, Guaranteed Policy, ICLR ’23, Naver Labs Europe, Grenoble, France