Unverified50% confidenceFactExact time
The study evaluated 40 native decision models and compared them with 10 general-purpose language models.
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
11/5/2026
First Seen
Valid until: 2/3/2027
Sources
Related Entities
Related Claims
UnverifiedThe research team evaluated more than ten representative models.70% similarUnverified研究表明在超过8-10轮对话后,模型对初始系统提示中规则的遵循率可下降20%-40%64% similarUnverifiedThe development team tested models with as many as 1 trillion parameters using the distributed layer-splitting method.63% similarVerifiedIn 2022, Hoffmann et al.'s Chinchilla paper redid experiments on four hundred models and concluded that, under the premise of compute-optimality, each parameter should correspond to approximately 20 tokens.62% similarUnverifiedIn a prompt injection defense evaluation, a certain decision model scored a perfect 40/40 on this test item.61% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/1137707API
curl https://kongchang.com/api/v1/knowledge/claims/1137707MCP
get_claim(id=1137707)