LLM Cache

Controls how the large language model (LLM) semantic cache stores and matches responses across your tenant.

The LLM Cache section of tenant configuration contains settings that were intended to control the behavior of Aisera's LLM semantic cache. These configurations are not currently implemented. Aisera stores the values but applies none of them at runtime.

Cache Length

Type	Integer
Default	1000

This configuration is deprecated and has no effect.

Cache Temperature

Type	Decimal
Default	0.1

This configuration is deprecated and has no effect.

Semantic Similarity Threshold

Type	Decimal
Default	0.95

This configuration is deprecated and has no effect.

Lower Semantic Similarity Threshold

Type	Decimal
Default	0.75

This configuration is deprecated and has no effect.

Keyword Similarity Threshold

Type	Decimal
Default	0.999

This configuration is deprecated and has no effect.