Scheduled Commit

data/retrieval_individual-8d38bb89-1e1a-471a-8b9e-35c1f784690e.jsonl

@@ -348,3 +348,5 @@
 {"tstamp": 1722379615.0159, "task_type": "retrieval", "type": "chat", "model": "voyage-multilingual-2", "gen_params": {}, "start": 1722379614.2951, "finish": 1722379615.0159, "ip": "", "conv_id": "81691119864942288cdc9471bbf81d7a", "model_name": "voyage-multilingual-2", "prompt": "who created the character spiderman?", "output": [["who created the character spiderman?", "Spider-Man\n\nSpider-Man is a superhero in American comic books published by Marvel Comics. Created by writer-editor Stan Lee and artist Steve Ditko, he first appeared in the anthology comic book Amazing Fantasy #15 (August 1962) in the Silver Age of Comic Books. He has been featured in comic books, television shows, films, video games, novels, and plays."]], "corpus": "wikipedia"}
 {"tstamp": 1722379655.2137, "task_type": "retrieval", "type": "chat", "model": "intfloat/e5-mistral-7b-instruct", "gen_params": {}, "start": 1722379654.9744, "finish": 1722379655.2137, "ip": "", "conv_id": "a2d40b7f31b54e1a9431e16ca8dfe52d", "model_name": "intfloat/e5-mistral-7b-instruct", "prompt": "A method for simulating specific individuals using large language models for role-playing.", "output": [["A method for simulating specific individuals using large language models for role-playing.", "Title: SimsChat: A Customisable Persona-Driven Role-Playing Agent\n\nAbstract: Large Language Models (LLMs) possess the remarkable capability to understand human instructions and generate high-quality text, enabling them to act as agents that simulate human behaviours. This capability allows LLMs to emulate human beings in a more advanced manner, beyond merely replicating simple human behaviours. However, there is a lack of exploring into leveraging LLMs to craft characters from several aspects. In this work, we introduce the Customisable Conversation Agent Framework, which employs LLMs to simulate real-world characters that can be freely customised according to different user preferences. The customisable framework is helpful for designing customisable characters and role-playing agents according to human's preferences. We first propose the SimsConv dataset, which comprises 68 different customised characters, 1,360 multi-turn role-playing dialogues, and encompasses 13,971 interaction dialogues in total. The characters are created from several real-world elements, such as career, aspiration, trait, and skill. Building on these foundations, we present SimsChat, a freely customisable role-playing agent. It incorporates different real-world scenes and topic-specific character interaction dialogues, simulating characters' life experiences in various scenarios and topic-specific interactions with specific emotions. Experimental results show that our proposed framework achieves desirable performance and provides helpful guideline for building better simulacra of human beings in the future. Our data and code are available at https://github.com/Bernard-Yang/SimsChat."]], "corpus": "arxiv"}
 {"tstamp": 1722379655.2137, "task_type": "retrieval", "type": "chat", "model": "embed-english-v3.0", "gen_params": {}, "start": 1722379654.9744, "finish": 1722379655.2137, "ip": "", "conv_id": "8d1338d4743f46caa513e773823a0f87", "model_name": "embed-english-v3.0", "prompt": "A method for simulating specific individuals using large language models for role-playing.", "output": [["A method for simulating specific individuals using large language models for role-playing.", "Title: Character-LLM: A Trainable Agent for Role-Playing\n\nAbstract: Large language models (LLMs) can be used to serve as agents to simulate human behaviors, given the powerful ability to understand human instructions and provide high-quality generated texts. Such ability stimulates us to wonder whether LLMs can simulate a person in a higher form than simple human behaviors. Therefore, we aim to train an agent with the profile, experience, and emotional states of a specific person instead of using limited prompts to instruct ChatGPT API. In this work, we introduce Character-LLM that teach LLMs to act as specific people such as Beethoven, Queen Cleopatra, Julius Caesar, etc. Our method focuses on editing profiles as experiences of a certain character and training models to be personal simulacra with these experiences. To assess the effectiveness of our approach, we build a test playground that interviews trained agents and evaluates whether the agents \\textit{memorize} their characters and experiences. Experimental results show interesting observations that help build future simulacra of humankind."]], "corpus": "arxiv"}
+{"tstamp": 1722379760.8385, "task_type": "retrieval", "type": "chat", "model": "text-embedding-3-large", "gen_params": {}, "start": 1722379760.1749, "finish": 1722379760.8385, "ip": "", "conv_id": "1732c4a6bebd4f80b51234d3c0c45e06", "model_name": "text-embedding-3-large", "prompt": "is there a second maze runner movie coming out", "output": [["is there a second maze runner movie coming out", "The Maze Runner (film)\n\nSequels On October 11, 2013, it was reported that Twentieth Century Fox had acquired the rights to the second book, The Scorch Trials. A screenplay was written by T. S. Nowlin, with director Wes Ball supervising the scriptwriting. The sequel was released on September 18, 2015. On July 25, 2014, Ball announced at San Diego Comic-Con that filming for the sequel would commence sometime between March and May 2015, should The Maze Runner become a success when it hits the theaters. However, two weeks prior to the film's release 20th Century Fox decided to move ahead with the sequel and pre-production began in early September 2014 in New Mexico. Cast members Dylan O'Brien, Kaya Scodelario, Thomas Brodie-Sangster, Ki-Hong Lee, and Patricia Clarkson reprised their roles for the sequel, as did director Wes Ball. It was announced that Aidan Gillen would be joining the film to play Janson (\"Rat-Man\"), as was Rosa Salazar who portrayed Brenda, Jacob Lofland who starred as Aris Jones, and Giancarlo Esposito who played Jorge Gallaraga.\nA second sequel, Maze Runner: The Death Cure was released on January 26, 2018.\nIn pop culture\nOn February 27, 2020, South Korean boy band BTS released their music video \"On\", which referenced set pieces from The Maze Runner. The homage was noted by source book author Dashner on Twitter.\nAnson Lo said the opening of his 2022 song \"King Kong\" music video paid tribute to The Maze Runner. He is the director of the music video."]], "corpus": "wikipedia"}
+{"tstamp": 1722379760.8385, "task_type": "retrieval", "type": "chat", "model": "text-embedding-004", "gen_params": {}, "start": 1722379760.1749, "finish": 1722379760.8385, "ip": "", "conv_id": "81025788b5004d79ab19447da56f0ff5", "model_name": "text-embedding-004", "prompt": "is there a second maze runner movie coming out", "output": [["is there a second maze runner movie coming out", "The Maze Runner (film)\n\nSequels On October 11, 2013, it was reported that Twentieth Century Fox had acquired the rights to the second book, The Scorch Trials. A screenplay was written by T. S. Nowlin, with director Wes Ball supervising the scriptwriting. The sequel was released on September 18, 2015. On July 25, 2014, Ball announced at San Diego Comic-Con that filming for the sequel would commence sometime between March and May 2015, should The Maze Runner become a success when it hits the theaters. However, two weeks prior to the film's release 20th Century Fox decided to move ahead with the sequel and pre-production began in early September 2014 in New Mexico. Cast members Dylan O'Brien, Kaya Scodelario, Thomas Brodie-Sangster, Ki-Hong Lee, and Patricia Clarkson reprised their roles for the sequel, as did director Wes Ball. It was announced that Aidan Gillen would be joining the film to play Janson (\"Rat-Man\"), as was Rosa Salazar who portrayed Brenda, Jacob Lofland who starred as Aris Jones, and Giancarlo Esposito who played Jorge Gallaraga.\nA second sequel, Maze Runner: The Death Cure was released on January 26, 2018.\nIn pop culture\nOn February 27, 2020, South Korean boy band BTS released their music video \"On\", which referenced set pieces from The Maze Runner. The homage was noted by source book author Dashner on Twitter.\nAnson Lo said the opening of his 2022 song \"King Kong\" music video paid tribute to The Maze Runner. He is the director of the music video."]], "corpus": "wikipedia"}