This letter could not be loaded. Try again in a moment.
These results are meant to be shared.
How it works. Position and topic classification is performed by a large language model (the open-weight Qwen3.8 27B) that reads each comment and labels its position — supports, opposes, or neutral — and the topics it raises. Copies and campaigns are found by hashes of the text and the overlap of six-to-eight-word phrases, grouped into text families; every unique figure counts one family once. Sentence embeddings (Qwen3) find reworded templates, shown under Organized campaigns, and cross-check the families; lexical, syntactic and stylometric features (TF-IDF, spaCy) score each comment’s originality for review. Neither decides a copy count. The full method is on the Methodologies page of roadless.org.
The full method. Every step, why it is done that way, its known limits and a glossary: Methodologies →
The data. The full canonical dataset — every comment on Docket FS-2025-0001 — is available from Regulations.gov: https://www.regulations.gov/bulkdownload.
Roadless.org — analysis of the public comment docket.
Media inquiries welcome — nicholas@wanderingnature.com.
© 2026 Roadless.org, Wandering Nature, LLC. All rights reserved.