Unverified50% confidenceFactExact time
Specification Gaming（规格利用）或Reward Hacking是强化学习对齐领域的经典问题，指模型学会利用评估程序漏洞而非真正解决底层问题
1
Sources
50%
Confidence
Long-term
Relevance
7/13/2026
First Seen
Sources
Related Claims
UnverifiedBenchmark reward hacking refers to the phenomenon where AI improves its scores by modifying the benchmark itself rather than genuinely optimizing the code.74% similarUnverifiedBenchmark hacking can artificially inflate benchmark scores through methods such as data contamination (test set leakage into training data), targeted training on the specific question formats of particular benchmarks, and selective reporting of the best benchmark results.67% similarUnverified评估器的质量决定整个系统的上限，设计不当可能导致奖励黑客现象67% similarUnverifiedBenchmark hacking refers to models potentially being optimized for specific test sets rather than possessing general capabilities, resulting in a gap between benchmark scores and real-world performance.67% similarUnverified在提示中包含明确验收标准的输出约束设计，本质上类似强化学习中奖励函数的作用64% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/500628API
curl https://kongchang.com/api/v1/knowledge/claims/500628MCP
get_claim(id=500628)