It makes me happy that the common machine-learning phrase "test set" is a palindrome.

👏👏👏 pham nguyen Rachael Sim qphong Zhiliang Chen Xinyuan Lin Xiaoqiang Xinyi Xu patrick jaillet Dai Zhongxiang Arun Verma Rui Qiao Wu Zhaoxuan Angel Pang Wei Koh Apivich H. (Kaotoo) Gregory Lau Bingchen Wang et al. Accepted ICLR 2026 #ICLR2025 AAAI #AAAI2025

Meta just dropped ReasonIR on Hugging Face Training Retrievers for Reasoning Tasks

When R1 came, I was thinking we should have a model trained to “reason” not only in English 🤔 Guess what, we show that with only English finetuning, the reasoning generalizes to other languages too! Models can also be “forced” to reason in other langs 🤯 However, more work

LLMs naturally memorize some verbatim of pre-training data. We study whether post-training can be an effective way to mitigate unintentional reproduction of pre-training data. 🛠️ No changes to pre-training or decoding 🔥 Training models to latently distinguish between memorized

Great work by Antoine Chaffin! It’s exciting to learn that training with the hard queries and hard negatives from our ReasonIR dataset can also significantly boost the performance of small models 😆 Looking forward to more advances in training efficient retrieval models!

🤯 We cracked RLVR with... Random Rewards?! Training Qwen2.5-Math-7B with our Spurious Rewards improved MATH-500 by: - Random rewards: +21% - Incorrect rewards: +25% - (FYI) Ground-truth rewards: + 28.8% How could this even work⁉️ Here's why: 🧵 Blogpost: tinyurl.com/spurious-rewar…

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.

Web data, the “fossil fuel of AI”, is being exhausted. What’s next?🤔 We propose Recycling the Web to break the data wall of pretraining via grounded synthetic data. It is more effective than standard data filtering methods, even with multi-epoch repeats! arxiv.org/abs/2506.04689

How cheerful it is to find a new-bloomed flower as all others have perished for several months in this winter time.😝