{"dep": 716, "title": "Provenance Erasure Rate A Compression-Survival Metric for Attribution Loss in AI-Composed Search Outputs", "date": "2026-05-03", "para": 86, "sent": 256, "sentence": "Judging LLM-as-a-judge with MT-Bench and Chatbot Arena. arXiv:2306.05685.", "alias": ["\\bchatbot arena\\b"], "cand_id": "D0001"}
{"dep": 741, "title": "The Evaluator Exists: Content-First Knowledge Assessment and the Political Economy of Proxy-Based Governance", "date": "2026-05-19", "para": 38, "sent": 80, "sentence": "(2023), in the MT-Bench / Chatbot Arena studies, established that GPT-4 reaches approximately 85% agreement with human preferences in controlled conversational-output evaluation, higher than the approximately 81% human-human agreement reported in the same studies.", "alias": ["\\bchatbot arena\\b"], "cand_id": "D0002"}
{"dep": 741, "title": "The Evaluator Exists: Content-First Knowledge Assessment and the Political Economy of Proxy-Based Governance", "date": "2026-05-19", "para": 202, "sent": 585, "sentence": "Judging LLM-as-a-Judge with MT-Bench and Chatbot Arena. arXiv:2306.05685. - Liang, W., Zhang, Y., Cao, H., Wang, B., Ding, D., Yang, X., Vodrahalli, K., He, S., Smith, D., Yin, Y., McFarland, D.", "alias": ["\\bchatbot arena\\b"], "cand_id": "D0003"}