Diagnosing and Correcting Reward Hacking in Verifiable Instruction Following
Paper • 2610.19525 • Published • 13
Kenneth Clark
CedarMap
AI & ML interests
None yet
Recent Activity
upvoted a paper 4 days ago
Diagnosing and Correcting Reward Hacking in Verifiable Instruction Following upvoted a paper 12 days ago
Semifactual Credit-Augmented Policy OptimizationOrganizations
None yet
upvoted a paper 4 days ago
upvoted a paper 9 days ago
upvoted a paper 12 days ago
upvoted 3
papers about 1 month ago
SpectralShift: Effective Context Window Extension of Gated DeltaNet via Spectral Reparameterization
Paper • 2609.14320 • Published • 34
Continual Learning Mechanisms Compose for Long-Horizon Memorization
Paper • 2609.06986 • Published • 369
Scaling Automatic Research Agents via World Models
Paper • 2608.12564 • Published • 481
upvoted 3
papers about 2 months ago
Knowing When Not to Reuse: Conditional Experience Transfer in Autonomous LLM Post-Training
Paper • 2608.26730 • Published • 133
LLaDA-Image: Building Strong Image Generators with Fully Open Training Recipes
Paper • 2609.03796 • Published • 116
PolicyGuide: From Guarding One Action to Guiding the Whole Workflow for Policy-Compliant LLM Agents
Paper • 2608.19861 • Published • 10
upvoted 5
papers 2 months ago
SemComp-Bench: Benchmarking Semantic Task Completion in Video Generation
Paper • 2608.17426 • Published • 160
FM-Bench: A Benchmark for Long-Horizon Management with Competing Agents
Paper • 2608.18423 • Published • 20
Can We Defend Against AI-Generated Video Attacks on Real-World Crisis Events? A Systematic Evaluation of Detectors, Generators and Social Dissemination
Paper • 2608.14391 • Published • 285
Specification-first convergence with an AI coding agent: a case study of dismantling a core architectural invariant across 189 files in a 717k-line codebase with no test oracle and no human code review
Paper • 2608.12440 • Published • 12
OpenART: Scaling Agent Red Teaming via Open-Ended Environment Evolution
Paper • 2608.00677 • Published • 263