Kunal Jha (@kjha02) 's Twitter Profile
Kunal Jha

@kjha02

CS PhD student @UW, prev. CSxPhilosophy @Dartmouth

ID: 1770229353001156608

linkhttp://kjha02.github.io calendar_today19-03-2024 23:22:36

36 Tweet

119 Followers

137 Following

Rui Xin (@rui_xin31) 's Twitter Profile Photo

Think PII scrubbing ensures privacy? 🤔Think again‼️ In our paper, for the first time on unstructured text, we show that you can re-identify over 70% of private information *after* scrubbing! It’s time to move beyond surface-level anonymization. #Privacy #NLProc 🔗🧵

Think PII scrubbing ensures privacy? 🤔Think again‼️ In our paper, for the first time on unstructured text, we show that you can re-identify over 70% of private information *after* scrubbing! It’s time to move beyond surface-level anonymization. #Privacy #NLProc 🔗🧵
Marlos C. Machado (@marloscmachado) 's Twitter Profile Photo

📢 I'm very excited to release AgarCL, a new evaluation platform for research in continual reinforcement learning‼️ Repo: github.com/machado-resear… Website: agarcl.github.io Preprint: arxiv.org/abs/2505.18347 Details below 👇

Max Kleiman-Weiner (@maxhkw) 's Twitter Profile Photo

LLMs learn beliefs and values from human data, influence our opinions, and then reabsorb those influenced beliefs, feeding them back to users again and again. We call this the "Lock-In Hypothesis" and develop theory, simulations, and empirics to test it in our latest ICML paper!

LLMs learn beliefs and values from human data, influence our opinions, and then reabsorb those influenced beliefs, feeding them back to users again and again. We call this the "Lock-In Hypothesis" and develop theory, simulations, and empirics to test it in our latest ICML paper!
Jacqueline He (@jcqln_h) 's Twitter Profile Photo

LMs often output answers that sound right but aren’t supported by input context. This is intrinsic hallucination: the generation of plausible, but unsupported content. We propose Precise Information Control (PIC): a task requiring LMs to ground only on given verifiable claims.

LMs often output answers that sound right but aren’t supported by input context. This is intrinsic hallucination: the generation of plausible, but unsupported content.

We propose Precise Information Control (PIC): a task requiring LMs to ground only on given verifiable claims.
Mickel Liu (@mickel_liu) 's Twitter Profile Photo

🤔Conventional LM safety alignment is reactive: find vulnerabilities→patch→repeat 🌟We propose 𝗼𝗻𝗹𝗶𝗻𝗲 𝐦𝐮𝐥𝐭𝐢-𝐚𝐠𝐞𝐧𝐭 𝗥𝗟 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 where Attacker & Defender self-play to co-evolve, finding diverse attacks and improving safety by up to 72% vs. RLHF 🧵

🤔Conventional LM safety alignment is reactive: find vulnerabilities→patch→repeat
🌟We propose 𝗼𝗻𝗹𝗶𝗻𝗲 𝐦𝐮𝐥𝐭𝐢-𝐚𝐠𝐞𝐧𝐭 𝗥𝗟 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 where Attacker & Defender self-play to co-evolve, finding diverse attacks and improving safety by up to 72% vs. RLHF 🧵
Kevin Ellis (@ellisk_kellis) 's Twitter Profile Photo

New paper: World models + Program synthesis by Wasu Top Piriyakulkij 1. World modeling on-the-fly by synthesizing programs w/ 4000+ lines of code 2. Learns new environments from minutes of experience 3. Positive score on Montezuma's Revenge 4. Compositional generalization to new environments