# language: en license: apache-2.0 library_name: htr tags: [text2text-generation, scientific-writing, style-reconstruction, research] pipeline_tag: text2text-generation

# HTR-Tiny-pilot-rc0 (v0.1.0)

Human Text Reconstruction Transformer — a research model that maps an LLM-rewritten scientific passage back towards the distribution of the pre-2018 human-authored original. Outputs are synthetic reconstructions: they are not the original text, and they are not evidence of human authorship or of AI use. Part of UQO working paper WP-AI-0002.

# Intended use

Research on how LLM rewriting transforms scientific prose; stylometric analysis; benchmarking inverse-paraphrase models; a tool for authors who want a draft re-aligned with historical scientific register as a suggestion that must be checked. Not intended for AI-text detection, authorship attribution, academic-integrity decisions, or production of text presented as human-written.

# How to use

from htr.inference.pipeline import HTRReconstructor
rec = HTRReconstructor.from_pretrained("path/to/HTR-Tiny-pilot-rc0")
r = rec.reconstruct("LLM-rewritten paragraph …", metadata={"discipline": "economics", "pub_year": 2012, "section_role": "results"})
print(r.reconstruction); print(r.warnings)

CLI: htr reconstruct --model path/to/HTR-Tiny-pilot-rc0 --text "…" --discipline economics --year 2012.

# Training data

Pilot: 5 383 validated Claude Opus 5.5 rewrite pairs of 974 PMC Open Access passages (CC0/CC BY/CC BY-NC/CC BY-NC-SA/public domain), after denoising pretraining on 17 k train-split PMC passages; research candidate, NOT the release model

# Evaluation (document-safe test split; identity baseline reported alongside)

{
  "chrf": {
    "mean": 63.6045,
    "ci": [
      62.1976,
      64.861
    ]
  },
  "edit_sim": {
    "mean": 0.5777,
    "ci": [
      0.5582,
      0.5955
    ]
  },
  "rougeL": {
    "mean": 0.5996,
    "ci": [
      0.5795,
      0.6169
    ]
  },
  "exact_match": {
    "mean": 0.0,
    "ci": [
      0.0,
      0.0
    ]
  },
  "number_recall": {
    "mean": 0.9308,
    "ci": [
      0.912,
      0.9503
    ]
  },
  "placeholders_exact": {
    "mean": 0.8845,
    "ci": [
      0.8303,
      0.9381
    ]
  },
  "_baseline_B0_identity": {
    "chrf": 71.2394,
    "edit_sim": 0.6665,
    "rougeL": 0.6898,
    "exact_match": 0.0,
    "number_recall": 1.0,
    "placeholders_exact": 1.0
  }
}

Metrics: chrF++, normalised edit similarity, ROUGE-L, exact match, number recall, placeholder (citation/math/URL) preservation, stylometric Wasserstein distance to originals. Confidence intervals are document-clustered bootstrap.

# Limitations and risks

• Reconstruction from a paraphrase is non-identifiable: many originals map to the same rewrite; the model returns one plausible estimate. Lexical overlap with the true original is bounded.
• Factual integrity is enforced heuristically (protected spans, number checks); residual errors are flagged in warnings, not guaranteed absent. Always verify numbers, citations and negations.
• Trained on Claude rewrites of CC-licensed biomedical, economics/finance, arXiv and public-domain book passages; behaviour on other rewriters, languages or genres is reported in the paper's generalisation experiments only.
• Memorisation: training originals were deduplicated and the model is evaluated for nearest-neighbour overlap; see the paper. Do not use to recover specific copyrighted texts.

# Licence

Code and weights: Apache-2.0. Data licences: see LICENSE-MODEL.txt and the data card. Attribution to NLM/PMC, arXiv, publishers and Project Gutenberg sources is in the data card.

# Citation

Boucher, S.-P. (2026). Human Text Reconstruction: Learning the Inverse of LLM-Induced Transformations in Scientific Writing. UQO Working Paper AI-0002.