Paper page - Hardening Agent Benchmarks with Adversarial Hacker-Fixer Loops
… View arXiv page View PDF GitHub 2 Add to collection Community Automatically hardening benchmarks and training environments with the hacker–fixer loop. …
Search
Paper page - Large Language Models Hack Rewards, and Society
… We hypothesise that the RL training process may exploit these gaps and therefore ask whether models' well-known tendency to hack reward functions during RL can scale into a more consequential failure mode named societal hacking: discovering loopholes in the rules society runs on. …
Paper page - Reproducing, Analyzing, and Detecting Reward Hacking in Rubric-Based Reinforcement Learning
… By injecting known biases into the LaaJ, CHERRL enables: Stable reproduction of reward hacking from a clean starting point Explicit observation of reward divergence between the biased and unbiased judges Precise identification of hacking onset step To demonstrate its utility, we analyze judge biase… …
Paper page - Do Coding Agents Deceive Us? Detecting and Preventing Cheating via Capped Evaluation with Randomized Tests
… The following papers were recommended by the Semantic Scholar API SpecBench: Measuring Reward Hacking in Long-Horizon Coding Agents 2026 Reward Hacking Benchmark: Measuring Exploits in LLM Agents with Tool Use 2026 Do Synthetic Trajectories Reflect Real Reward Hacking? …
Paper page - G-Zero: Self-Play for Open-Ended Generation from Zero Data
… View arXiv page View PDF GitHub 22 Add to collection Community Self-evolving LLMs excel in verifiable domains but struggle in open-ended tasks, where reliance on proxy LLM judges introduces capability bottlenecks and reward hacking. …
VibeGame: Exploring Vibe Coding Games
… If you have feedback or feature requests the Roblox MCP server, feel free to submit issues on the project page: https://github.com/Roblox/studio-rust-mcp-server Also it is open source, so you can hack on it and add more context management tools here: https://github.com/Roblox/studio-rust-mcp-server… …
Paper page - Representation over Routing: Diagnosing Temporal Routing Pathologies in Multi-Timescale PPO
… Code and reproducible scripts are open-sourced in the repo. the core idea that really sticks is target decoupling: keep multi-timescale predictions on the critic for auxiliary representation learning, while the actor updates are driven only by long-horizon advantages. this separation seems to block… …
Paper page - EurekAgent: Agent Environment Engineering is All You Need For Autonomous Scientific Discovery
… View arXiv page View PDF GitHub 65 Add to collection Community As model capabilities continue to improve, we argue that the bottleneck for autonomous scientific discovery is shifting from prescribing agent workflows to designing agent environments: the resources, constraints, and interfaces that sh… …
Paper page - The Meta-Agent Challenge: Are Current Agents Capable of Autonomous Agent Development?
… To ensure evaluation integrity, this framework is secured by multi-layer defenses against reward hacking . Leveraging this framework, we demonstrate that meta-agent s rarely match human-engineered baseline policies, and the few that do are dominated by proprietary frontier models. …
Paper page - SAAS: Self-Aware Reinforcement Learning for Over-Search Mitigation in Agentic Search
… View arXiv page View PDF GitHub 7 Add to collection Community Agentic search enables LLMs to solve complex multi-hop questions through iterative reasoning and external search. …