Hi, I’m Kaushik. I work on making AI systems more interpretable and controllable.
Visiting Researcher @ University of Cambridge; MSc Computer Science @ ETH Zürich
Recent writing
Selected projects
Characterised goalkeeper actions from skeletal data by learning robust representations via contrastive learning and analysing their structure using …
Studied LLaVA-1.5-7B using a controlled synthetic benchmark; revealed an existence bias causing yes-bias in binary relational tasks through logit …
Adapted Apertus 8B/70B for Swiss legal summarisation, showing full fine-tuning outperforms GPT-4o and Claude 3.5 Sonnet on BERTScore and ROUGE, while …