Alpha
Contact
Support · Boldt-DC

Will it last?

Ownership levelPartialnone·limited·partial·substantial·fullAnalytical input D ยท 45.2/100

This page is a projection of the one entry record, the Transparency factor that Support covers. The full verdict is set by all four factors together, floor-weighted so the weakest caps the whole.

Which domain expands which factor
  • AssessUse & modify + Transparency
  • ImplementData control + Reliability
  • UseReliability
  • SupportTransparency

Common problems & fixes

  • Poor or off-topic output usually means a chat template was applied. Use plain completion.
  • Load the tokenizer from the same repo as the weights. Each Boldt model has its own 32k tokenizer.
  • The Boldt-DC-1B sample loads Boldt/Boldt-1B. Replace the name with Boldt/Boldt-DC-1B.
  • The 350M card says "three models" and lists four, and calls the model "Boldt-DC-300M" in its table. The repo, the organisation page and the API all say 350M, and the config counts 468M parameters (ev-card-350m, ev-hf-platform).
  • The 1B card carries a custom_code tag, but the repo holds no custom code.

Versions, changelog & cadence

Both repos were created on 2026-05-02. The weights have not changed since the initial upload; later commits edit the README (the last on 2026-11-01 for the 350M repo). The organisation added 15 research checkpoints and three classifiers on 2026-08-30, and the paper appeared in EMNLP 2026 proceedings. There is no changelog file, so pin the commit hash (ev-hf-platform, ev-paper).

Community & support channels

The Hugging Face discussions tab on both repos is open and empty. The chair's project page and the paper list the authors' university addresses, which are not a security channel (ev-hf-platform, ev-project).

Tracked known issues

There is no public issue tracker. The publisher lists the missing toxicity and bias evaluation and the risk of bias amplification from repeated epochs (ev-card-1b, ev-paper). The card errors above are the other known defects.

How this scores

The ownership factor this domain covers, drawn from the one entry record.

2

TransparencyDo you know what it is: weights, training, behaviour, and legible terms?

Moderate

Weights, training dataset, filter classifiers and a detailed paper are public, but training code, configs, the label set and an evaluation harness are not, so the models cannot be rebuilt exactly (ev-paper, ev-dataset).

How this scores (AOI sub-dimensions)
Provenance2/5how well we can trace and verify what went into the modelSafetensors only, Hugging Face LFS hashes, a scan that flagged nothing and plain Llama-architecture files with no custom code give about 4 of 8 checklist boxes.
Governance3/5how accountable and well-documented the publisher isAn accountable university chair with named authors, a peer-reviewed paper (EMNLP 2026), a project page, two release batches (May and August 2026) and README edits as late as 2026-11-01.
What this means for adoptionYou own the use of the weights: Apache-2.0 is declared, ungated, commercial use is allowed and your data stays on your hardware. You do not own a deployable German assistant, because these are small base models with no safety evaluation, you cannot rebuild them without the unreleased training code, and every benchmark is the publisher's own. The level rises if the group ships a LICENSE file, releases its training code and configs, and settles the Llama 3.3 question on the classifiers.

Sources

The same evidence records as the entry sheet. Read means the text was verified; unverified means it is known to exist but not yet read.

Model cardread2026-11-19
The Boldt-DC-350M card declares Apache-2.0 in its metadata only, says the Boldt series is 'trained from scratch' (a publisher claim), calls the model a base model not optimised for chat, and says no systematic toxicity, bias or stereotype evaluation was done.
Model cardread2026-11-19
The Boldt-DC-1B card declares Apache-2.0 in its metadata only, gives a 2,048-token context and 200B training tokens of the German Dense-Core subset, reports a 44.05 average on the publisher's German benchmark suite, and carries the same no-safety-evaluation statement.
Licenceread2026-11-19
Both repos declare 'license: apache-2.0' in front matter and are not gated.
Technical_reportread2026-11-19
The EMNLP 2026 paper 'Repetition over Diversity' (arXiv 2604.28075, Aynetdinov, Haller and Akbik, Humboldt-Universitaet zu Berlin) describes training models 'from scratch' on filtered German FineWeb-2.
Documentationread2026-11-19
Boldt/fineweb2-de-dc is a public ODC-By 1.0 dataset of 24.4M German documents (28B tokens) subject to Common Crawl's Terms of Use, selected by three released classifiers trained on Llama-3.3-70B-Instruct labels; the classifier repos have no README, no licence and no LICENSE file.
Terms of serviceread2026-11-19
Llama 3.3 Community License section 1.b.i says that anyone who uses Llama Materials or any outputs of them to train or improve a distributed AI model shall include 'Llama' at the beginning of that model's name.
Documentationread2026-11-19
The Boldt organisation is not verified on Hugging Face and has two members.
Model cardread2026-11-19
Sibling releases differ from the DC models.
Vendor announcementread2026-11-19
The Hugging Face organisation README and the project page name the Chair of Machine Learning at Humboldt-Universitaet zu Berlin as the developer and describe Boldt as 'state-of-the-art' and 'top-tier'; these are marketing statements.
Third-party analysisread2026-11-19
No independent benchmark, leaderboard entry or safety report for any Boldt model was found; the only outside artefacts are community fine-tunes on the Hugging Face Hub.