ParlAI Vote is an interactive system for exploring European Parliament debates and votes, testing Large Language Models (LLMs) for vote prediction and bias analysis. The platform links debate topics, speeches, and list voting results, including rich demographic data such as gender, age, country, and political group. Users can explore debates, review linked speeches, compare state-of-the-art LLM predictions with actual voting results, and view error breakdowns by demographic group. By visualizing the EuroParlVote benchmark and the core tasks of gender classification and vote prediction, ParlAI Vote highlights systematic performance biases in state-of-the-art LLMs. The system integrates data, models, and visual analytics into a single interface, lowering the barriers to result replication, behavioral audits, and running counter-empirical scenarios. It supports research, education, and public engagement in legislative decision-making, while also highlighting the strengths and limitations of current LLMs in political analysis.
