Unverified50% confidenceBenchmarkExact time
Large language models perform extremely well in single-turn scenarios, but their performance degrades significantly when the same complete information is split across multiple turns and provided incrementally.
1
Sources
50%
Confidence
Long-term
Relevance
10/6/2026
First Seen
Sources
Related Entities
Related Claims
Unverified大语言模型多卡张量并行时，每一层前向传播都需要在卡间同步激活值，带宽瓶颈会直接拉低整体推理吞吐77% similarUnverifiedTraditional large language models often fail at complex tasks spanning multiple steps and multiple time points due to context loss, errors in intermediate steps, or goal drift.76% similarUnverifiedThe larger the parameters of a large language model, the higher the computational requirements for a single forward inference pass, and latency increases accordingly.75% similarUnverifiedLarge language models are prone to hallucination drift when handling detail consistency problems that require precise long-range context tracking.74% similarUnverifiedLarge language models expose their weaknesses when multi-step reasoning, spatial imagination, or creative thinking outside the distribution of training data is required.74% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/982134API
curl https://kongchang.com/api/v1/knowledge/claims/982134MCP
get_claim(id=982134)