The first user-centric benchmark grounded in real user feedback signals for evaluating preference alignment and dialogue generation is presented, demonstrating that user feedback prediction is a learnable capability and revealing different aspects that influence user experience.
Mengze Hong, Xia Zeng, Zeyang Lei et al.· arXiv.org· 1 citation
This work ventures deep into detokenization using activation patching in controlled paired experiments that isolate the contribution of different model components, localizing English detokenization in Llama2-7B to a two-stage process at Layer 1.
ReadingMachine is a computational methodology for structured corpus reading that uses large language models to perform bounded reading operations over entire document collections. Rather than relying on retrieval or recursive summarization, the approach decomposes analysis into inspectable stages including insight extr...
Large language model (LLM) agents increasingly rely on external experience to continually adapt to changing environments without modifying their underlying models. Recent experience mechanisms have demonstrated promising results across diverse tasks. However, their effectiveness is typically evaluated within individual...
Yining Zhu, Zihao Deng, Leiming Wang et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
DySem is proposed, a novel training-free framework that investigates more semantic-related internal components of LLMs via multilingual consensus, and shifts away from static representation spaces in favor of dynamic, sample-specific semantic dimensions by constructing text-dependent joint semantic set and computes sim...
Kaijie Zheng, Wei-Qin Wang, Yi-Le Wang et al.· arXiv.org· 0 citations
CorVer (Corpus Verify) is a lightweight training-time approach to process supervision that derives sentence-level rewards from corpus co-occurrence statistics and outperforms the unmodified models in all standard factual-QA settings.
Shi-Chen Fan, Hao-Chang Hao, De-Hai Min et al.· 5 citations
ResearchMath-14k is introduced, a set of research-level mathematical problems curated from academic sources via a multi-agent pipeline that ranks above existing math datasets on knowledge, novelty, and procedural difficulty and suggests that research-level mathematical content and greater reasoning diversity may help e...
Experimental results across a broad range of domains show that MCPO alleviates performance degradation caused by mixed-domain training and outperforms single-domain training in most cases.
Zong-Ji Yu, Wen-Shui Luo, Yi-Liu Sun et al.· arXiv.org· 3 citations
This work introduces HiMed, a Hindi reasoning medical corpus and benchmark suite covering both Western and Indian medicine and proposes HiMed-8B, a Hindi-form medical reasoning LLM, through the design of decaying scaffolding reward.
Dingfeng Jiang, Handi Yan, Chen Ma et al.· arXiv.org· 0 citations
A search-based memory framework called BOOKMARKS for active grounding, which retains access to the full preceding storyline and collects task-relevant information on demand, and improves next-action fidelity in a five-dimensional evaluation.
As Large Language Models (LLMs) are increasingly deployed in cross-linguistic contexts, ensuring safety across diverse regulatory and cultural environments has become a critical challenge. However, existing multilingual benchmarks largely rely on general risk taxonomies and machine-translated data, limiting evaluation...
Yunhan Zhao, Zhaorun Chen, Xingjun Ma et al.· 0 citations
We identify and systematically characterize a class of task-structural alignment failures in large language models (LLMs): even when the harmful intent remains unchanged, changing the task presentation and output constraints can substantially alter model safety behavior. Specifically, when a harmful request is reformul...
Yuheng Chen, Zhiyu Wu, Bowen Cheng et al.· 0 citations
Flint-interactive gives AI agents a language for creating interactive visualizations. Click a chart, filter data, highlight related content, and explore beyond the view the author originally chose. The post Flint-interactive: A language for creating interactive data visualizations appeared first on Microsoft Research.
A new computational approach reveals how cells communicate to assemble into tissues. This could accelerate stem cell engineering, improve disease models, and inform therapies.
Microsoft Research Blog· microsoft.comOct 15, 2026
Zero-knowledge proofs can verify computing results without rerunning workloads or exposing sensitive data. Magnetar helps bring them into the mainstream with support for standard programming languages, major speed gains, and quantum-resistant security. The post Proof-based general-purpose computing with Magnetar appeared first on Microsoft Research.
MIT News · Artificial Intelligence· news.mit.eduSep 24, 2026
