Postdoc @UNC working on NLP, AI, and computational linguistics. Formerly PhD student @JHU and undergrad @McGill esteng.github.iohttps://bsky.app/profile/esteng.bsky.socialhttps://bsky.app/profile/esteng.bsky.social/post/3loh7f7djwc2iExtremely excited to announce that I will be joining @utaustin.bsky.social Computer Science in August 2025 as an Assistant Professor! 🎉05 May 2025 20:28 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3loh7f7djwc2ihttps://bsky.app/profile/esteng.bsky.social/post/3lo22ypikkc2x🌵 I'm going to be presenting PBT at #NAACL2025 today at 2PM! Come by poster session 2 if you want to hear about: -- balancing positive and negative persuasion -- improving LLM teamwork/debate -- training models on simulated dialogues With @mohitbansal.bsky.social and @peterbhase.bsky.social [contains quote post or other embedded content]30 Apr 2025 15:04 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lo22ypikkc2xhttps://bsky.app/profile/esteng.bsky.social/post/3lnxtvspvxs2n✈️ Heading to #NAACL2025 to present 3 main conf. papers, covering training LLMs to balance accepting and rejecting persuasion, multi-agent refinement for more faithful generation, and adaptively addressing varying knowledge conflict. Reach out if you want to chat!29 Apr 2025 17:52 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lnxtvspvxs2nhttps://bsky.app/profile/esteng.bsky.social/post/3lnkypmdp2s26Check out 🚨CAPTURe🚨 -- a new benchmark testing spatial reasoning by making VLMs count objects under occlusion. SOTA VLMs (GPT-4o, Qwen2-VL, Intern-VL2) have high error rates on CAPTURe (but humans have low error ✅) and models struggle to reason about occluded objects. arxiv.org/abs/2504.15485 🧵👇24 Apr 2025 15:14 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lnkypmdp2s26https://bsky.app/profile/esteng.bsky.social/post/3lmmq2pgiv22g🚨Announcing TaCQ 🚨 a new mixed-precision quantization method that identifies critical weights to preserve. We integrate key ideas from circuit discovery, model editing, and input attribution to improve low-bit quant., w/ 96% 16-bit acc. at 3.1 avg bits (~6x compression) 📃 arxiv.org/abs/2504.0738912 Apr 2025 14:19 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lmmq2pgiv22ghttps://bsky.app/profile/esteng.bsky.social/post/3lixt5nvtkc2d🚨UPCORE is our new method for balancing unlearning/forgetting with maintaining model performance. Best part is it works by selecting a coreset from the data rather than changing the model, so it is compatible with any unlearning method, with consistent gains for 3 methods + 2 tasks! [contains quote post or other embedded content]25 Feb 2025 02:33 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lixt5nvtkc2dhttps://bsky.app/profile/esteng.bsky.social/post/3lherazzd3c2a🚨 Excited to announce UTGen and UTDebug, where we first learn to generate unit tests and then apply them to debugging generated code with LLMs, with strong gains (+12% pass@1) on LLM-based debugging across multiple models/datasets via inf.-time scaling and cross-validation+backtracking! 🧵👇 [contains quote post or other embedded content]04 Feb 2025 19:13 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lherazzd3c2ahttps://bsky.app/profile/esteng.bsky.social/post/3lggdownyzk2i🎉Very excited that our work on Persuasion-Balanced Training has been accepted to #NAACL2025! We introduce a multi-agent tree-based method for teaching models to balance: 1️⃣ Accepting persuasion when it helps 2️⃣ Resisting persuasion when it hurts (e.g. misinformation) arxiv.org/abs/2410.14596 🧵 1/423 Jan 2025 16:50 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lggdownyzk2ihttps://bsky.app/profile/esteng.bsky.social/post/3lgbn2gpvzc2gCongrats @mohitbansal.bsky.social for being selected to be part of this prestigious #AAAI Fellows group! Very well-deserved recognition of long-term contributions 🎉🎉 [contains quote post or other embedded content]21 Jan 2025 19:55 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lgbn2gpvzc2ghttps://bsky.app/profile/esteng.bsky.social/post/3lfueqw7ttk2zCongrats on this huge (and well-deserved) accomplishment @mohitbansal.bsky.social! [contains quote post or other embedded content]16 Jan 2025 13:21 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lfueqw7ttk2zhttps://bsky.app/profile/esteng.bsky.social/post/3le7mxbdb722dApply soon to - work with tons of extremely talented and strong students and get hands-on experience mentoring - benefit from expert mentorship - be in a great department and live in an area with high quality of life and more! Feel free to ping me if you want to hear more from a current postdoc! [contains quote post or other embedded content]26 Dec 2024 13:57 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3le7mxbdb722dhttps://bsky.app/profile/esteng.bsky.social/post/3lcybai6ebs2j✈️ I've landed in Vancouver for #NeurIPS2024 11/12: LACIE, a pragmatic speaker-listener method for training LLMs to express calibrated confidence: arxiv.org/abs/2405.21028 12/12: GTBench, a benchmark for game-theoretic abilities in LLMs: arxiv.org/abs/2402.12348 P.s. I'm on the faculty market👇 [contains quote post or other embedded content]10 Dec 2024 22:14 +0000at://did:plc:drvtnoosdjozr7dmv5a4stwt/app.bsky.feed.post/3lcybai6ebs2j