Backend
What RL on Expert-Built Tasks Teaches Nemotron 3.5 Super
arXiv cs.SE0 views
arXiv:2611.22161v1 Announce Type: cross Abstract: We post-train NVIDIA's Nemotron 3.5 Super (120B total, 12B active parameters) with reinforcement learning (RL) on five expert-built task datasets, one at a time: real-world software engineering, complex constraints, long-horizon tool use, professional chart understanding and enterprise document reasoning. Each run trains a fresh rank-32 LoRA adapter with CISPO on Tinker, with no supervised stage, and each adapter is compared with the base model on the same tasks, harness and settings. On twelve headline benchmarks, eight built by other groups and four by Surge AI, the gains range from +4.5 points on IFBench to +13.8 on GDP_pdf, which rises from 10.4% to 24.2%; SWE-bench Pro V2 rises from 86.9% to 92.1%, Terminal-Bench 2.1 from 63.9% to 70.3%, MultiChallenge from 58.3% to 66.6%, AutomationBench from 14.8% to 27.5% and Chartography from 39.4% to 50.6%. The document adapter, trained without API tools, also gains +8.0 points on AutomationBench. In attempts read side by side with the base model's on tasks each adapter newly solves, the base model's failures recur in a few modes, and each adapter does more often what the base model missed: the tool-use adapter finds the data and the rules before acting, the coding adapter writes its own test or scratch files on SWE-bench Pro V2, the constraints adapter carries earlier turns into its answers, the chart adapter measures the marks in code, and the document adapter reads AutomationBench's spreadsheets.
Source:https://arxiv.org/abs/2611.22161PreviousAdaptive Budget Allocation for Smart-Contract Fuzzing: Why Online Feedback Fails to Improve Vulnerability TriageNextBeyond Bitwidth: Deployment-Constrained Mixed-Precision Quantization for Multilingual LLM Inference on NVIDIA Blackwell
Related posts
Driving the Style: Velocity-Conditioned Whole-Body Control from Motion Demonstrations
0arXiv cs.RO
sqrtVINS-Based 360 Visual-Inertial SLAM Systems: Hilti-Trimble SLAM Challenge 2026 Entry
0arXiv cs.RO
Ontology-Compiled Control Barrier Functions for Heterogeneous Robot Teams
0arXiv cs.RO
I Used to Get Attached to My Code. I Don't Anymore.
0dev.to AI
Securing MCP Servers: Preventing Shadow Tool Declarations and JSON-RPC Parameter Poisoning
0dev.to AI
How do I find out why users abandon conversations with my AI agent?
0dev.to AI