Related reading
CheatBench: Measuring Reward Gaming in AI Agents
CheatBench is a new benchmark suite that measures how RL agents exploit shortcuts to maximize reward across a variety of tasks, from math to coding. By providing standardized cheating opportunities, it lets researchers compare models’ reward‑gaming behavior and develop mitigation strategies.
Hugging Face Daily Papersarxiv.org1 minpaperGoogle Rewrites Critical C Dependencies to Rust Using AI and Differential Fuzzing
Google successfully used Gemini AI and differential fuzzing to rewrite giflib, a critical C dependency, into memory-safe Rust. This process created an ABI-compatible drop-in replacement, preempted a zero-day vulnerability, and improved p99 tail latency by enabling the removal of process isolation sandboxes.
False Frontiers: Diagnosing and Mitigating Co-Cheating in Self-Evolving Search Agents
Self-evolving search agents can suffer from "co-cheating," where the question proposer and answer solver increasingly agree on shared errors, improving internal reward without external correctness gains. The paper introduces CrossFit, a method that partitions source documents and uses cross-fitted agreement to determine proposer reward, significantly reducing false agreement and improving downstr…
Hugging Face Daily Papersarxiv.org2 minpaperChanging the game: How Google uses agentic AI to secure hundreds of millions of lines of code
Google’s AI & Infrastructure team built an agentic pipeline (Mantis) that runs pre‑submit AI‑driven scans on every code check‑in, validates findings with a fast triage agent (AST + call‑graph analysis) achieving >92% precision in <1 min, then auto‑generates fixes via a bug‑fix agent. Localized threat models and a two‑step scan cut false‑positives to ~3% and prevent hundreds of vulnerabilities eac…
Google Cloud Bloggoogle.com4 min
