{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,7]],"date-time":"2026-03-07T18:45:15Z","timestamp":1772909115612,"version":"3.50.1"},"reference-count":15,"publisher":"IEEE","license":[{"start":{"date-parts":[[2024,12,15]],"date-time":"2024-12-15T00:00:00Z","timestamp":1734220800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2024,12,15]],"date-time":"2024-12-15T00:00:00Z","timestamp":1734220800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,12,15]]},"DOI":"10.1109\/bigdata62323.2024.10825845","type":"proceedings-article","created":{"date-parts":[[2025,1,16]],"date-time":"2025-01-16T18:31:23Z","timestamp":1737052283000},"page":"8805-8807","source":"Crossref","is-referenced-by-count":3,"title":["Emerging Trends in LLM Benchmarking"],"prefix":"10.1109","author":[{"given":"Akshar Prabhu","family":"Desai","sequence":"first","affiliation":[]},{"given":"Ritu","family":"Prajapati","sequence":"additional","affiliation":[]},{"given":"Tejasvi","family":"Ravi","sequence":"additional","affiliation":[]},{"given":"Mohammad","family":"Luqman","sequence":"additional","affiliation":[]},{"given":"Pranjul","family":"Yadav","sequence":"additional","affiliation":[]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1145\/3641289"},{"key":"ref2","volume-title":"Holistic evaluation of language models","author":"Liang","year":"2023"},{"key":"ref3","volume-title":"Livebench: A challenging, contamination-free llm benchmark","author":"White","year":"2024"},{"key":"ref4","volume-title":"A generalist agent","author":"Reed","year":"2022"},{"key":"ref5","volume-title":"Beyond the imitation game: Quantifying and extrapolating the capabilities of language models","author":"Srivastava","year":"2023"},{"key":"ref6","volume-title":"Osworld: Benchmarking multimodal agents for open-ended tasks in real computer environments","author":"Xie","year":"2024"},{"key":"ref7","volume-title":"Evaluating large language models: A comprehensive survey","author":"Guo","year":"2023"},{"key":"ref8","volume-title":"Logiqa: A challenge dataset for machine reading comprehension with logical reasoning","author":"Liu","year":"2020"},{"key":"ref9","volume-title":"Reclor: A reading comprehension dataset requiring logical reasoning","author":"Yu","year":"2020"},{"key":"ref10","volume-title":"Program synthesis with large language models","author":"Austin","year":"2021"},{"key":"ref11","volume-title":"Humaneval on latest gpt models \u2013 2024","author":"Li","year":"2024"},{"key":"ref12","article-title":"What disease does this patient have? A large-scale open domain question answering dataset from medical exams","volume-title":"CoRR","author":"Jin","year":"2020"},{"key":"ref13","article-title":"Truthfulqa: Measuring how models mimic human falsehoods","volume-title":"CoRR","author":"Lin","year":"2021"},{"key":"ref14","first-page":"3008","article-title":"Learning to summarize with human feedback","volume-title":"Advances in Neural Information Processing Systems","volume":"33","author":"Stiennon","year":"2020"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.findings-emnlp.301"}],"event":{"name":"2024 IEEE International Conference on Big Data (BigData)","location":"Washington, DC, USA","start":{"date-parts":[[2024,12,15]]},"end":{"date-parts":[[2024,12,18]]}},"container-title":["2024 IEEE International Conference on Big Data (BigData)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10824975\/10824942\/10825845.pdf?arnumber=10825845","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,1,17]],"date-time":"2025-01-17T08:36:54Z","timestamp":1737103014000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10825845\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,12,15]]},"references-count":15,"URL":"https:\/\/doi.org\/10.1109\/bigdata62323.2024.10825845","relation":{},"subject":[],"published":{"date-parts":[[2024,12,15]]}}}