Get started
Author

Shashwat Saxena

Observed name: Saxena, Shashwat
Observed institution: Carnegie Mellon University
OpenAlex ID: A5120404554
Terminal Wrench: A Dataset of 331 Reward-Hackable Environments and 3,632 Exploit Trajectories
2026 · Bercovich, Ivan, Segal, Ivgeni, et al. · arXiv (Cornell University) · 16
PDF
Hardening Agent Benchmarks with Adversarial Hacker-Fixer Loops
2026 · Zhong, Ziqian, Segal, Ivgeni, et al. · arXiv (Cornell University) · 14
PDF
OGPO: Sample Efficient Full-Finetuning of Generative Control Policies
2026 · Patil, Sarvesh, Nakamoto, Mitsuhiko, et al. · arXiv (Cornell University) · 5
PDF
Jailbreaking in the Haystack
2025 · Shah, Rishi Rajesh, Wu, Chen Henry, et al. · arXiv (Cornell University) · 5
PDF
Hodoscope: Unsupervised Monitoring for AI Misbehaviors
2026 · Zhong, Ziqian, Saxena, Shashwat, et al. · arXiv (Cornell University) · 4
PDF
Scalable Weak-to-Strong Verifier Hardening for Agent Environments
2026 · Zeng, Chong, Zhong, Ziqian, et al. · arXiv (Cornell University) · 1
PDF
Reinforcement Learning Enables Scaling Generalist Robot Policy Improvement
2026 · Korniak, Michal, Bagatella, Marco, et al. · arXiv (Cornell University)
PDF
VaDar: Implicit Goal Conditioning via Value Disaggregation
2026 · Saxena, Shashwat, Goel, Mehul, et al. · arXiv (Cornell University)
PDF
Works count: 8
Citation count: 45
H-index: 4
I10-index: 2
Year
Topic
Type