dsup2026/plumbline-0.6b-reader

Plumbline-0.6B reader

The second adapter of Plumbline-0.6B, which judges whether a text breaks a rule written in plain English. The reader works on the rule alone and reads two facts from its wording, so that the main card's cutoffs can be chosen with nothing for you to label:

• does a text break the rule by including something it forbids, or by leaving something out;
• can a text keep the rule without it ever coming up, as a dish without nuts keeps a rule about nuts.

On 160 rules written blind it reads the first fact right for 94% of them and the second for 95%.

Use it

It sits beside the judging adapter on the same base model, and the snippet in the main card loads it:

model.load_adapter("dsup2026/plumbline-0.6b-reader", adapter_name="reader")

Synthetic data generated by a Qwen model (Yang et al., 2025) was used in training.

Licence

PolyForm Noncommercial 1.0.0. Commercial use needs a separate licence.

The adapter is under PolyForm Noncommercial 1.0.0, reproduced in full in LICENSE. Any noncommercial purpose is permitted with no permission needed, which its own text spells out as personal study, private entertainment, hobby projects and amateur pursuits, and which covers charities, schools, public research bodies and government.

Commercial use needs a separate licence. Contact me via the details in LICENSE. Running the model over a group or a queue belonging to a business, a client or an employer is commercial use however informally it is run and whoever owns the laptop.

The base model remains Apache-2.0 and is not relicensed here; retain its LICENSE and NOTICE.

Citation

@misc{qwen3technicalreport,
  title         = {Qwen3 Technical Report},
  author        = {Yang, An and Li, Anfeng and Yang, Baosong and others},
  year          = {2025},
  eprint        = {2505.09388},
  archivePrefix = {arXiv},
  primaryClass  = {cs.CL},
  url           = {https://arxiv.org/abs/2505.09388},
  note          = {A Qwen model generated the synthetic training data; the base model is Qwen3-Reranker-0.6B}
}