Category
3,240 papers
Small models are made more capable through distillation from a larger one that shares their tokenization scheme. However, do distilled byte and token models behave similarly in terms of scaling trends as compute and data increases? To enable this comparison, we introduce two variants to efficiently convert token logits...
This paper presents an open and modular transformer-based framework that detects and structures social tipping point evidence at the passage level, outperforming both a climate-focused model and untuned language models.
The results indicate that front-loading code understanding into a reusable, codebase-specialised vector store enables small local models to deliver grounded and repository-specific answers, making the agent well suited as a privacy-preserving development tool for in-house scientific codebases.
Assistants built on large language models are expected to write as their user would, and the dominant approach is single-channel: preferences summarised from conversation history and reinserted into context. This inverts the order of inference. Preferences are the task-dependent surface of a comparatively stable person...
Reach audiences
Advertise in front of researchers, engineers, and readers.
Blinded physician evaluation has been considered by many to be the gold standard for assessing clinical reasoning in large language models (LLMs). This is difficult to scale; thus, prior studies typically rely on small physician panels, often from a single institution or specialty, which both limits the scientific ques...
The Graph Theory Agent (GTA), which pairs a preference-trained representation selector with plan-and-decompose scaffolding around a frozen executor LLM, is proposed, which lifts Phi-4 from 53.5% to 69.1% on the benchmark's easy split and from 33.0% to 41.5% on its hard split.
This work asks whether a self-architecting transformer, which grows its own width and depth from quantitative signals derived from the attention matrices, can extract competitive protein representations from a single reference proteome.
CARE-MoE is proposed, an efficient MoE LLM inference framework comprising two core components that balances expert placement by jointly modeling co-activation correlation and hot–cold drift, preventing overload from correlated experts and enabling low-cost adaptive rebalancing.
MIGServe treats the physical layout of MIG instances as a first-class scheduling dimension through three techniques: buddy-aware partition placement, which preserves large contiguous free blocks by allocating next to existing occupied buddies; proactive pair-matching migration, which consolidates fragmented half-full b...
A Multi-State Semantic Cache (MSSC) framework that dynamically maps knowledge chunks into three mutually exclusive states: zero cache, index cache, and full cache is proposed that reduces average service latency by over 40% in dynamic scenarios and suppresses backhaul reconfiguration traffic by up to 38% compared to co...
A sail is pushed by the wind, so a boat cannot outrun the wind──so it seems. This paper shows that changing only the heading removes the cap──neither the wind nor the sail is altered at all. No new theorem or law is claimed. Scope of this paper (scope note): No new theorem or law is claimed──apparent wind, propulsion b...
The Hall--Petch relation says finer grains are stronger, and it reverses at 10 to 30 nm. This paper shows that the reversal is not because the theoretical strength was reached──the prediction there does not even reach 14 per cent of it. No new theorem or law is claimed. Scope of this paper (scope note): No new theorem...
A new computational approach reveals how cells communicate to assemble into tissues. This could accelerate stem cell engineering, improve disease models, and inform therapies.
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.
Learn how Google DeepMind and GovTech Singapore uses lightweight on-device Gemma models to intercept conversational scams in private chats with zero cloud data egress.
StudentSim is a research system for evaluating and improving AI tutoring tools before they are tested with learners. It models how individuals may engage with tutors, giving feedback to improve tutor performance across subjects like chess, writing, and mathematics. The post Simulating student learning for better AI tutoring appeared first on Microsoft Research.
We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.
