1 source filed · RU 1
Nine classes (parties, document details, contracts, object of dispute, performance, payments, breaches, damages, correspondence), each extracted by a separate call returning JSON.
0.70 vs 0.71macro-F1 of the distilled Qwen2.5-1.5B vs Ministral 3 14B on a 145-document golden set
1 source filed · KO 1
Tasks are identified by hashing the start of the system prompt, where role and task definition stay stable even when few-shot examples change.
about 33-37%daily share of product-matching requests the small model handled in about 6 days of production
2 sources filed · EN 2
Summarising reasoning already hurt distillers: a model trained on full Opus 4.8 transcripts scored 80.4% on SWE-bench Verified, against 71.8% on summaries.
around 90 to 99%claimed cut in data harvested by distillers
Zero-shot small models failed: five off-the-shelf SLMs missed the quality bar, and base Qwen3-8B produced malformed tool calls on 23% of cases.
0.913agreement of a LoRA-tuned Qwen3-8B with recorded production decisions