Technical report · EN
Context Rot: How Increasing Input Tokens Impacts LLM Performance
Chroma technical report on how increasing input token counts affects model performance.
Open resource ↗Recommended reading
Selected research and technical resources for exploring models and their capabilities.
3 resources
Research paper · EN
Intelligence Degradation in Long-Context LLMs: Critical Threshold Determination via Natural Length Distribution Analysis
Analysis of degradation in long-context language models; the requested version v1 link is preserved.
Open resource ↗
Research paper · EN
Lost in the Middle: How Language Models Use Long Contexts
Study of how information position within context affects language models' use of it.
Open resource ↗
No resources match your search.