Hello! I’m a Staff Research Scientist and Tech Lead at Google Research, where I build factually grounded Large Language Models. My focus is on making Gemini both more knowledgeable and less prone to hallucinations. My research spans the full development cycle of LLMs, from designing pre-training strategies to reinforcement learning techniques that help models better distinguish fact from fiction.
I’m fascinated with a fundamental question: how do language models actually ‘know’ things? My research interests center on how LLMs acquire, represent, and apply knowledge to reason about the world. I’m particularly interested in the mechanisms by which they internalize factual information during pre-training and how that knowledge can be reliably extracted to enable reasoning.
I completed a PhD in Natural Language Processing at Tel Aviv University, supported by a Google PhD Fellowship. My doctoral research, as well as my three internships at Google, focused on reasoning over structured data. Prior to my PhD, I was a Research Staff Member at IBM Research, where I worked on deep learning for language understanding. My research has been recognized with best paper awards at INLG and NAACL.
Feel free to reach out if you are interested in collaborating!
📜 Publications
| arXiv 2025 | Gemini 2.5: Pushing the Frontier with Advanced Reasoning, Multimodality, Long Context, and Next Generation Agentic Capabilities Gemini Team, ..., Jonathan Herzig and others  |
| COLM 2025 | Inside-out: Hidden factual knowledge in LLMs Zorik Gekhman, Eyal Ben David, Hadas Orgad, Eran Ofek, Yonatan Belinkov, Idan Szpektor, Jonathan Herzig, Roi Reichart |
| arXiv 2025 | DRAGged into Conflicts: Detecting and Addressing Conflicting Sources in Search-Augmented LLMs Arie Cattan, Alon Jacovi, Ori Ram, Jonathan Herzig, Roee Aharoni, Sasha Goldshtein, Eran Ofek, Idan Szpektor, Avi Caciularu  |
| Findings of ACL 2024 | Multilingual instruction tuning with just a pinch of multilinguality Uri Shaham, Jonathan Herzig, Roee Aharoni, Idan Szpektor, Reut Tsarfaty, Matan Eyal |
| ACL 2024 | A Chain-of-Thought Is as Strong as Its Weakest Link: A Benchmark for Verifiers of Reasoning Chains Alon Jacovi, Yonatan Bitton, Bernd Bohnet, Jonathan Herzig, Or Honovich, Michael Tseng, Michael Collins, Roee Aharoni, Mor Geva  |
| arXiv 2024 | Constructing benchmarks and interventions for combating hallucinations in llms Adi Simhi, Jonathan Herzig, Idan Szpektor, Yonatan Belinkov |
| EMNLP 2024 | Does Fine-Tuning LLMs on New Knowledge Encourage Hallucinations? Zorik Gekhman, Gal Yona, Roee Aharoni, Matan Eyal, Amir Feder, Roi Reichart, Jonathan Herzig |
| NeurIPS 2024 | TACT: Advancing Complex Aggregative Reasoning with Information Extraction Tools Avi Caciularu, Alon Jacovi, Eyal Ben-David, Sasha Goldshtein, Tal Schuster, Jonathan Herzig, Gal Elidan, Amir Globerson |
| ICML 2024 | Representation Surgery: Theory and Practice of Affine Steering Shashwat Singh, Shauli Ravfogel, Jonathan Herzig, Roee Aharoni, Ryan Cotterell, Ponnurangam Kumaraguru |
| arXiv 2024 | Can few-shot work in long-context? recycling the context to generate demonstrations Arie Cattan, Alon Jacovi, Alex Fabrikant, Jonathan Herzig, Roee Aharoni, Hannah Rashkin, Dror Marcus, Avinatan Hassidim, Yossi Matias, Idan Szpektor, Avi Caciularu |
| arXiv 2024 | Distinguishing ignorance from error in llm hallucinations Adi Simhi, Jonathan Herzig, Idan Szpektor, Yonatan Belinkov |
| Findings of ACL 2023 | mFACE: Multilingual Summarization with Factual Consistency Evaluation Roee Aharoni, Shashi Narayan, Joshua Maynez, Jonathan Herzig, Elizabeth Clark, Mirella Lapata |
| NeurIPS 2023 | What You See is What You Read? Improving Text-Image Alignment Evaluation Michal Yarom, Yonatan Bitton, Soravit Changpinyo, Roee Aharoni, Jonathan Herzig, Oran Lang, Eran Ofek, Idan Szpektor |
| EMNLP 2023 | Trueteacher: Learning factual consistency evaluation with large language models Zorik Gekhman, Jonathan Herzig, Roee Aharoni, Chen Elkind, Idan Szpektor |
| EMNLP 2023 | Evaluating and Modeling Attribution for Cross-Lingual Question Answering Benjamin Muller, John Wieting, Jonathan H Clark, Tom Kwiatkowski, Sebastian Ruder, Livio Baldini Soares, Roee Aharoni, Jonathan Herzig, Xinyi Wang |
| GEM Workshop 2023 | QAMPARI: A benchmark for open-domain questions with many answers Samuel Amouyal, Tomer Wolfson, Ohad Rubin, Ori Yoran, Jonathan Herzig, Jonathan Berant |
| Findings of EMNLP 2023 | A Comprehensive Evaluation of Tool-Assisted Generation Strategies Alon Jacovi, Avi Caciularu, Jonathan Herzig, Roee Aharoni, Bernd Bohnet, Mor Geva |
| Compendium of Neurosymbolic Artificial Intelligence 2023 | Latent Trees for Compositional Generalization Jonathan Herzig, Jonathan Berant, Ben Bogin |
| NAACL 2022 | Learning To Retrieve Prompts for In-Context Learning Ohad Rubin, Jonathan Herzig, Jonathan Berant |
| NAACL 2022 | TRUE: Re-evaluating Factual Consistency Evaluation Or Honovich, Roee Aharoni, Jonathan Herzig, Hagai Taitelbaum, Doron Kukliansy, Vered Cohen, Thomas Scialom, Idan Szpektor, Avinatan Hassidim, Yossi Matias |
| EMNLP 2022 | Evaluating the Impact of Model Scale for Compositional Generalization in Semantic Parsing Linlu Qiu, Peter Shaw, Panupong Pasupat, Tianze Shi, Jonathan Herzig, Emily Pitler, Fei Sha, Kristina Toutanova |
| arXiv 2022 | Attributed question answering: Evaluation and modeling for attributed large language models Bernd Bohnet, Vinh Q Tran, Pat Verga, Roee Aharoni, Daniel Andor, Livio Baldini Soares, Massimiliano Ciaramita, Jacob Eisenstein, Kuzman Ganchev, Jonathan Herzig, Kai Hui, Tom Kwiatkowski, Ji Ma, Jianmo Ni, Lierni Sestorain Saralegui, Tal Schuster, William W. Cohen, Michael Collins, Dipanjan Das, Donald Metzler, Slav Petrov, Kellie Webster |
| ACL 2021 | Span-based semantic parsing for compositional generalization Jonathan Herzig, Jonathan Berant |
| NAACL 2021 | Open Domain Question Answering over Tables via Dense Retrieval Jonathan Herzig, Thomas Müller, Syrine Krichene, Julian Martin Eisenschlos |
| arXiv 2021 | Unlocking compositional generalization in pre-trained models using intermediate representations Jonathan Herzig, Peter Shaw, Ming-Wei Chang, Kelvin Guu, Panupong Pasupat, Yuan Zhang |
| EMNLP 2021 | Finding needles in a haystack: Sampling Structurally-diverse Training Sets from Synthetic Data for Compositional Generalization Inbar Oren, Jonathan Herzig, Jonathan Berant |
| ACL 2020 | TAPAS: Weakly Supervised Table Parsing via Pre-training Jonathan Herzig, Paweł Nowak, Thomas Müller, Francesco Piccinno, Julian Martin Eisenschlos |
| Findings of EMNLP 2020 | Improving Compositional Generalization in Semantic Parsing Inbar Oren, Jonathan Herzig*, Nitish Gupta*, Matt Gardner, Jonathan Berant |
| NAACL 2019 | CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledge  Best Resource Paper Award Alon Talmor*, Jonathan Herzig*, Nicholas Lourie, Jonathan Berant |
| NAACL 2019 | Value-based Search in Execution Space for Mapping Instructions to Programs Dor Muhlgay, Jonathan Herzig, Jonathan Berant |
| ACL 2019 | TalkSumm: A Dataset and Scalable Annotation Method for Scientific Paper Summarization Based on Conference Talks Guy Lev, Michal Shmueli-Scheuer, Jonathan Herzig, Achiya Jerbi, David Konopnicki |
| *SEM 2019 | Bot2vec: Learning representations of chatbots Jonathan Herzig, Tommy Sandbank, Michal Shmueli-Scheuer, David Konopnicki |
| UMAP 2019 | Detecting persuasive arguments based on author-reader personality traits and their interaction Michal Shmueli-Scheuer, Jonathan Herzig, David Konopnicki, Tommy Sandbank |
| EMNLP 2019 | Don't paraphrase, detect! Rapid and Effective Data Collection for Semantic Parsing Jonathan Herzig, Jonathan Berant |
| EMNLP 2019 | A Summarization System for Scientific Documents Shai Erera, Michal Shmueli-Scheuer, Guy Feigenblat, Ora Peled Nakash, Odellia Boni, Haggai Roitman, Doron Cohen, Bar Weiner, Yosi Mass, Or Rivlin, Guy Lev, Achiya Jerbi, Jonathan Herzig, Yufang Hou, Charles Jochim, Martin Gleize, Francesca Bonin, Debasis Ganguly, David Konopnicki |
| EMNLP 2018 | Decoupling Structure and Lexicon for Zero-Shot Semantic Parsing Jonathan Herzig, Jonathan Berant |
| NAACL 2018 | Detecting Egregious Conversations between Customers and Virtual Agents Tommy Sandbank, Michal Shmueli-Scheuer, David Konopnicki, Jonathan Herzig, John Richards, David Piorkowski |
| IUI 2018 | On the Expression of Agent Emotions in Customer Support Dialogs in Social Media Michal Shmueli-Scheuer, Jonathan Herzig, Tommy Sandbank, David Konopnicki |
| ACL 2017 | Neural semantic parsing over multiple knowledge-bases Jonathan Herzig, Jonathan Berant |
| IUI 2017 | Ehctool: Managing emotional hotspots for conversational agents Tommy Sandbank, Michal Shmueli-Scheuer, Jonathan Herzig, David Konopnicki, Rottem Shaul |
| INLG 2017 | Neural response generation for customer service based on personality traits Best Short Paper Award Jonathan Herzig, Michal Shmueli-Scheuer, Tommy Sandbank, David Konopnicki |
| ICTIR 2017 | Emotion detection from text via ensemble classification using word embeddings Jonathan Herzig, Michal Shmueli-Scheuer, David Konopnicki |
| CSCW 2016 | I understand your frustration Guy Feigenblat, David Konopnicki, Michal Shmueli-Scheuer, Jonathan Herzig, Hen Shkedi |
| UMAP 2016 | Predicting customer satisfaction in customer support conversations in social media using affective features Jonathan Herzig, Guy Feigenblat, Michal Shmueli-Scheuer, David Konopnicki, Anat Rafaeli |
| SIGDial 2016 | Classifying emotions in customer support dialogues in social media Jonathan Herzig, Guy Feigenblat, Michal Shmueli-Scheuer, David Konopnicki, Anat Rafaeli, Daniel Altman, David Spivak |
| IEEE Transactions on Biomedical Engineering 2014 | Monitoring cardiac stress using features extracted from S1 heart sounds Jonathan Herzig, Amitai Bickel, Arie Eitan, Nathan Intrator |
| Hypertext 2014 | An author-reader influence model for detecting topic-based influencers in social media Jonathan Herzig, Yosi Mass, Haggai Roitman |
| ImmersiveMe 2014 | Mindful: A Platform for Large-Scale Affective Field Research Guy Feigenblat, Jonathan Herzig, Michal Shmueli-Scheuer, David Konopnicki |
| IBM Journal of Research and Development 2013 | A statistical approach to mining customers' conversational data from social media David Konopnicki, Michal Shmueli-Scheuer, Doron Cohen, Benjamin Sznajder, Jonathan Herzig, Ariel Raviv, N Zwerling, Haggai Roitman, Yosi Mass |
| PCI 2013 | Eventsense: Capturing the pulse of large-scale events by mining social media streams Emmanouil Schinas, Symeon Papadopoulos, Sotiris Diplaris, Yiannis Kompatsiaris, Yosi Mass, Jonathan Herzig, Lazaros Boudakidis |