Unverified50% confidenceBenchmarkExact time
In 16 multi-turn real human conversation tests, 2.0 outperformed the base model comprehensively with a score of 16 to 0.
1
Sources
50%
Confidence
Long-term
Relevance
10/2/2026
First Seen
Sources
Related Entities
Related Claims
UnverifiedAnthropic分析100万条真实对话后发现，人类提问水平和AI回答水平的相关性系数达到0.9265% similarUnverifiedThe 2023 Microsoft Research paper 'LIMA: Less Is More for Alignment' showed that using only 1,000 carefully curated high-quality samples can enable a base model to exhibit conversational capabilities comparable to fine-tuning on tens of thousands of samples.64% similarUnverifiedThe same ordering—relation information preceding entity information—is observed across all 16 model–threshold combinations (threshold range 0.2 to 0.5).63% similarUnverifiedThat experiment reached the adaptation criterion specified in the paper (three consecutive sessions passing) at session 14, with both the bold rate and list rate metrics showing a downward trend across the first 36 sessions.62% similarUnverifiedAlthough full-modal large models support voice input and output, their scores on agent benchmarks are far lower than those of mainstream reasoning models.62% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/968171API
curl https://kongchang.com/api/v1/knowledge/claims/968171MCP
get_claim(id=968171)