Hakan Inan
RESEARCH SCIENTIST | MENLO PARK, UNITED STATES
Hakan Inan is a research scientist at Meta, working at the intersection of generative AI and trust & safety, currently focusing on large language models (LLM) safety. He has also done work in topics in natural language processing including conversational assistants, Natural Language Understanding & Generation, and Machine Translation, as well as topics in trust & safety, including AI systems to improve safety of real-time interactions in the Metaverse. Hakan holds a PhD from Stanford University, with a thesis on machine learning and computational neuroscience.
Research Areas
Hakan's Publications
December 07, 2023
CONVERSATIONAL AI
NLP
Llama Guard: LLM-based Input-Output Safeguard for Human-AI Conversations
Hakan Inan, Davide Testuggine, Jianfeng Chi, Kartikeya Upasani, Krithika Iyer, Madian Khabsa, Rashi Rungta, Yuning Mao
December 07, 2023
July 18, 2023
CONVERSATIONAL AI
NLP
Llama 2: Open Foundation and Fine-Tuned Chat Models
Moya Chen, Marie-Anne Lachaux, Thibaut Lavril, Hugo Touvron, Adina Williams, Alan Schelten, Amjad Almahairi, Andrew Kuan, Andrew Poulton, Angela Fan, Anthony Hartshorn, Artem Korenev, Aurelien Rodriguez, Binh Tang, Brian Fuller, Cristian Canton Ferrer, Cynthia Gao, Dan Bikel, David Esiobu, Diana Liskovich, Xiaoqing Ellen Tan, Eric Michael Smith, Guillem Cucurull, Hakan Inan, Igor Molybog, Iliyan Zarov, Isabel Kloumann, Puxin Xu, Jenya Lee, Jeremy Fu, Jeremy Reizenstein, Jude Fernandes, Kalyan Saladi, Kevin Stone, Louis Martin, Lukas Blecher, Madian Khabsa, Marcin Kardas, Melanie Kambadur, Naman Goyal, Nikolay Bashlykov, Peter Albert, Praj Bhargava, Punit Singh Koura, Pushkar Mishra, Ranjan Subramanian, Rashi Rungta, Robert Stojnic, Ross Taylor, Ruan Silva, Rui Hou, Saghar Hosseini, Sergey Edunov, Sharan Narang, Shruti Bhosale, Soumya Batra, Thomas Scialom, Todor Mihaylov, Vedanuj Goswami, Viktor Kerkez, Wenyin Fu, Xavier Martinet, Yasmine Babaei, Yinghai Lu, Yixin Nie, Yuchen Zhang, Yuning Mao, Zheng Yan
July 18, 2023
Our approach
Latest news
Foundational models
Meta © 2026