Benchmark

DICE-300

A benchmark for evaluating the function-calling capability of large language models, comprising 300 test cases.

Timeline (last 90 days)

Oct 19

On the Qwen3.5-9B model, SchemaShield-Lite raised accuracy on the DICE-300 benchmark from 0.717 to 0.813, recovering 19 of the 22 drift-induced failures.

Unverified50%

All Facts (1)

Unverified

On the Qwen3.5-9B model, SchemaShield-Lite raised accuracy on the DICE-300 benchmark from 0.717 to 0.813, recovering 19 of the 22 drift-induced failures.

50%

Source Articles

SchemaShield-Lite: Fixing LLM Agent Interface Drift Without RetrainingOct 19