会员服务
Between Spoilers and Guesswork: Avoiding Over-Specification in Coding Benchmarks
Arxiv
0+阅读 · 10月16日
Multi-Reward Hacking: From Aggregation-Induced Drift to Trust-Anchored Defense
Arxiv
0+阅读 · 10月15日
Rubric Dropout: A Simple Way to Mitigate Reward Hacking in Rubric-as-Reward RL
Arxiv
0+阅读 · 10月13日
ESAU: Expeditious Swarm Remote Attestation for Pub-Sub IoT Networks with Untrusted Cloud-based Proxy Verifier
Arxiv
0+阅读 · 10月14日
Who Sets the Pace of Judge Hacking? Judge Capability versus the Policy-Judge Gap
Arxiv
0+阅读 · 10月14日
Rent-Seeking Language Models: Reward Models Set the Prices, Frequent Audits Curb the Rent
Arxiv
0+阅读 · 10月14日
Diagnosing and Correcting Reward Hacking in Verifiable Instruction Following
Arxiv
0+阅读 · 10月14日
Gradient Routing Localizes and Removes Unintended Behaviors in RL
Arxiv
0+阅读 · 10月12日
Who Sets the Pace of Judge Hacking? Judge Capability versus the Policy-Judge Gap
Arxiv
0+阅读 · 10月11日
UMM-RM: An Upcycle-and-Merge MoE Reward Model for Mitigating Reward Hacking
Arxiv
0+阅读 · 10月11日
Copying Is Not Grounding: Training Faithful Long-Form RAG with Authored-Sentence Credit
Arxiv
0+阅读 · 10月13日
Reward Hacking as a Transmissible Strategy in Multi-Agent Coding Systems
Arxiv
0+阅读 · 10月12日
SAGA: Score-Weighted Adaptive Generation Alignment for Low-Resource Nordic Language Models
Arxiv
0+阅读 · 10月13日
Adversarial Reward Auditing for Active Detection and Mitigation of Reward Hacking
Arxiv
0+阅读 · 9月28日
How Hackable Is Your Speech Quality Metric? A Corrected Protocol, a Benchmark, and What Patching Buys
Arxiv
0+阅读 · 10月7日
参考链接
微信扫码咨询专知VIP会员
Top