ToolGrad: Efficient tool-use dataset generation with textual "gradients"
Machine Intelligence
More from the blog
GPT-Lab
· gpt-lab.eu
Sep 17, 2026
Beyond Prompt Engineering: The Role of Tacit Knowledge in Software Engineering
AI is making software generation faster, but speed does not remove the need for expertise. As more work is delegated to AI, tacit knowledge may become one of the most important human advantages in software engineering. The post Beyond Prompt Engineering: The Role of Tacit Knowledge in Software Engineering appeared first on GPT-Lab.
New AI technique could make minimally invasive surgeries safer and more precise
This patient-specific method, called xvr, helps doctors use X-rays for surgical navigation in fields such as orthopedics and neurosurgery.
Measure by measure, studying society accurately
Naoki Egami has become a standout in political methodology, helping refine tools that give scholars durable results.
Called to serve: Tech, research, and positive impact with Chris White
Lab Director Chris White has worked on research challenges with real-world implications—from new approaches to wartime data analysis to tools for combating human trafficking. He talks to program manager Weishung Liu about the influences that led to the work and more. The post Called to serve: Tech, research, and positive impact with Chris White appeared first on Microsoft Research.
Related papers
Computational and AI-Driven Ecosystem for Structure-Based Covalent Drug Discovery.
This Account describes a computational and AI-driven ecosystem for structure-based covalent drug discovery and dives into a suite of cutting-edge, AI-driven computational methods, exploring the potential of deep learning in tasks such as molecular docking, covalent binding site prediction, and lead optimization.
ChargeNet: E(3) Equivariant Graph Attention Network for Atomic Charge Prediction
This work introduces an advanced equivariant graph attention neural network specifically engineered to model long-range atomic electrostatic interactions with high precision, and improves the model's accuracy, generalization, and robustness in complex scenarios.
Constitutional Midtraining: Content Presence Drives Alignment Gains
Post-training alignment is often shallow, eroding under fine-tuning. It remains untested as to whether constitutional midtraining interventions can produce durable alignment when cleanly isolated from post-training. We build a 394M-token constitutional corpus from Anthropic's Constitution and apply constitutional midtraining at 120B scale, where principled, values-based content is inserted into midtraining. A 2x2 design (curriculum ordering x deliberative reasoning) was used to produce four constitutionally midtrained conditions, plus a control, which were evaluated on self-generated and established benchmarks including alignment under pressure, value conflict resolution, blackmail, and emergent misalignment. All models were evaluated across three stages: post-midtraining, post-SFT, and post-benign fine-tuning. Constitutionally midtrained models outperformed the control on alignment generalization and durability, notably on blackmail: SFT instilled a blackmail propensity in all models, but constitutional midtraining blunted it, with the advantage surviving benign fine-tuning (-17.5pp). This durability did not extend to settings that required active resistance to in-context pressure or conflict, where the advantage attenuates after SFT. The presence of constitutional content at midtraining also mattered more than its structure, and constitutional midtraining incurred no capability cost, on average, at any stage (MMLU, ARC-Easy, piqa, GSM8K). A modest amount of constitutional content at midtraining could therefore yield broad, persistent alignment gains, offering a cheap, complementary addition to SFT-centered pipelines. Code, data, and models are available.
Let the Flows Tell: Solving Graph Combinatorial Optimization Problems with GFlowNets
This paper designs Markov decision processes (MDPs) for different combinatorial problems and proposes to train conditional GFlowNets to sample from the solution space and demonstrates that GFlowNet policies can efficiently find high-quality solutions.