OPEN
ANT decision model
ANT adds trained decision heads to frozen Qwen backbones. Read its synthetic tests, validation-fitted escalation and separate backbone downloads.
By Semantique Team · Hugging Face weight release: 10 Nov 2026
Trained candidate-scoring heads for frozen Qwen3.5-0.8B and 4B backbones. Each package contains the head, temperature calibration and an execute-or-escalate policy, with a separate backbone download. The English/Italian training and evaluation situations are synthetic.
Quick answers
What is ANT by Semantique Team?
ANT: Trained candidate-scoring heads for frozen Qwen3.5-0.8B and 4B backbones.
Who makes ANT?
ANT: Semantique Team
Facts
- Package
- Head-only releases; ANT-08B has 1,053,696 trainable parameters and a 4.2 MB head, excluding the backbone
- Inputs
- Text state, question and supplied candidates; configured context budget 1,024 tokens and 128 per candidate
- Runtime
- Python runtime from the matching ant-08b or ant-4b GitHub repository; optional MLX path
- Status
- Work in progress; one training run per size, without measured seed variance
- Licence
- Apache-2.0 for released heads; Qwen backbones retain their own Apache-2.0 licence
Claims and sources
- The maker reports test accuracy 67.44% for ANT-08B and 73.92% for ANT-4B on 648 synthetic situations. On 456 situations from held-out use cases, accuracy is 57.02% and 59.65%. These are separate from the validation-fitted escalation policy. vendor reportedsource
- ANT-08B reports validation coverage 44.2% at 9.9% error among accepted decisions, using 688 validation rows. Its fitted threshold does not establish that error rate on new domains; OOD ECE is 0.220. vendor reportedsource
Links
Source: ANT-4B card → Source: ANT-08B card → Source: 4B metrics → Source: 08B policy → Source: Licence →
Related projects
No related projects are listed in the registry.