Yuqing Yang

🧐When do LLMs admit their mistakes when they should know better? In our new paper, we define this behavior as retraction: the model indicates that its generated answer was wrong. LLMs can retract—but they rarely do.🤯 arxiv.org/abs/2505.16170 👇🧵

A practical insight!

Announcing 🔭✨Hubble, a suite of open-source LLMs to advance the study of memorization! Pretrained models up to 8B params, with controlled insertion of texts (e.g., book passages, biographies, test sets, and more!) designed to emulate key memorization risks 🧵

# 🚨 4B open-recipe model beats Claude-4-Opus 🔓 100% open data, recipe, model weights and code. Introducing Polaris✨--a post-training recipe for scaling RL on advanced reasoning models. 🥳 Check out how we boost open-recipe reasoning models to incredible performance levels (65 → 79 on AIME25) through RL training on open-source data and academic-level resources. 📑Notion: honorable-payment-890.notion.site/POLARIS-A-POst… 📗Blog post: hkunlp.github.io/blog/2025/Pola… 🤗Model & data: huggingface.co/POLARIS-Project 💻Code: github.com/ChenxinAn-fdu/…

There’s been hot debate about (The Illusion of) The Illusion of Thinking. My take: it’s not that models can’t reason — they just aren’t perfect at long-form generation yet. We eval reasoning models on LongProc benchmark (requiring generating 8K CoTs, see thread). Reasoning models actually outperform instruction-tuned ones, showing the benefit of long CoT training for long outputs. Still, there’s plenty of room for improvement. See details in our post 👉 princeton-pli.github.io/LongProc/rmode…

🧵 Recent studies show LLMs can self-improve their responses when given external feedback. But how effectively can they incorporate it? We tested this systematically—and found they can't fully integrate feedback, even when the feedback is high-quality and backed by ground-truth.

🚨 We discovered a surprising side effect of Reinforcement Finetuning (RFT): it makes LLMs more confidently wrong on unanswerable questions. We call this the hallucination tax: a drop in refusal behavior that leads to overconfident hallucinations. 🧵 1/n

Textual steering vectors can improve visual understanding in multimodal LLMs! You can extract steering vectors via any interpretability toolkit you like -- SAEs, MeanShift, Probes -- and apply them to image or text tokens (or both) of Multimodal LLMs. And They Steer!

🧵7/7 📄Paper: arxiv.org/abs/2505.16170 💻 Code & data: github.com/ayyyq/llm-retr… 🙌 Many thanks to my awesome advisor @robinomial and everyone at Allegro Lab!

🧵6/7 In short: Retraction is belief-driven. And belief can be probed, steered, and trained. We hope our work contributes to a better understanding of self-correction in LLMs.

🧐When do LLMs admit their mistakes when they should know better? In our new paper, we define this behavior as retraction: the model indicates that its generated answer was wrong. LLMs can retract—but they rarely do.🤯 arxiv.org/abs/2505.16170 👇🧵

Want to know what your LLM don’t know? This is how 👇 Preprint: arxiv.org/abs/2503.23361 Code: github.com/uscnlp-lime/SEA

Billion-parameter LLMs still struggle with simple arithmetic? 📞 FoNE (Fourier Number Embedding) tackles this problem. By mapping numbers directly into Fourier space, it bypasses tokenization and significantly improves numerical accuracy with better efficiency and accuracy.

Running your model on multiple GPUs but often found the speed not satisfiable? We introduce Ladder-residual, a parallelism-aware architecture modification that makes 70B Llama with tensor parallelism ~30% faster! Work done at @togethercompute. Co-1st author with @MayankMish98 and mentored by @ben_athi, @tri_dao arxiv.org/pdf/2501.06589 🧵[1/7]

Come join the #NeurIPS2024 poster session and discuss whether language models can learn to skip steps in reasoning! 🗓Dec 12, Thursday, 11:00 am - 2:00 pm 📍East Exhibit Hall A-C #2900 Feel free to stop by and say hi! I am actively seeking Summer 2025 internship opportunities!

@ShuaichenChang Thank you!!

@yyqcode We (AWS AI LAB) are hiring!

Arrived in Vancouver for #NeurIPS2024 🇨🇦! I'll be presenting Alignment for Honesty, a year-old paper that still fascinates me with how LLMs navigate knowledge boundaries. Also glad to chat about self-correction and reasoning. Actively seeking a 2025 summer internship!