Unverified50% confidenceProductExact time
Livenerf 是一个开源项目,用于持续追踪监测Claude Opus模型是否被削弱(nerf/降智)
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/30/2026
First Seen
Valid until: 12/29/2026
Sources
Related Entities
Related Claims
Unverified为应对测试环境应试化,研究者建议引入模型无法预知的隐藏测试、动态评估、活基准(live benchmark)等方法63% similarUnverifiedKubernetes中的Liveness Probe和Readiness Probe是心跳检测思想在容器编排领域的工程实现60% similarUnverifiedBrad Abrams conducted a live demo using Claude Opus 4 to demonstrate A/B test analysis capabilities59% similarVerifiedLiveCodeBench 引入了时间戳过滤策略,筛选模型训练截止日期之后发布的题目,以确保评测数据对被测模型而言是全新的56% similarUnverifiedEvidence Core 的副标题为「Build live analytics with coding agents」55% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/960970API
curl https://kongchang.com/api/v1/knowledge/claims/960970MCP
get_claim(id=960970)