TAO $223.70+2.6% 24h
119

CapabilityForge Next

sn119Audio & SpeechWatch48Avoid20

CapabilityForge Next rewards traceable training-data improvements that measurably enhance the accuracy, reasoning, clarity, age appropriateness, and safety of learning-oriented AI through reproducible evaluation.

Emission
0
paused on chain
Alpha
0.0108
τ · mcap 17.6K
Top-slot payout
τ per winning epoch · live
Stars
0
live from the GitHub API
Primary language
repo-reported
Last push
7d ago
feeds the dormancy integrity signal
Topics
repo-declared tags

About the repo

what the project says about itself — the input for semantic labels

capabilityforge-next/capabilityforge· pushed 7d ago

CapabilityForge Next rewards traceable training-data improvements that measurably enhance the accuracy, reasoning, clarity, age appropriateness, and safety of learning-oriented AI through reproducible evaluation.

Reading this tab

scope of the data

Taonets fetches the repo's public metadata (description, topics, language, stars, last push) and the head of its README. Commit-count histories, contributor lists and release notes require the GitHub commits/releases endpoints, which the worker doesn't consume yet — those panels are coming with the ingest upgrade and are not simulated in the meantime.

README head

first lines of the default branch README, unedited

# ![intro](./images/intro.jpg)CapabilityForge Next

> Verifiable data improvements for better learning intelligence.

**CapabilityForge Next** is Bittensor Subnet 119—an incentive network for discovering training-data improvements that make learning-oriented AI more accurate, clear, age-appropriate, and safe.

The subnet initially focuses on mathematics and the natural sciences, where answers, reasoning steps, calculations, and explanations can be evaluated objectively. Its scope can expand to additional subjects, languages, and learner profiles as reliable evaluation standards become available.

## How It Works

- **Miners** generate bounded synthetic or appropriately licensed data patches with record-level provenance.
- **Validators** check data rights, privacy, lineage, duplication, contamination, and evaluation leakage.
- Qualified patches are tested against a frozen baseline under identical training and evaluation conditions.
- Rewards reflect measurable capability improvements—not dataset size, compute expenditure, or unverified claims.

```text
GENERATE → VERIFY → TRAIN → EVALUATE → REWARD
```

## What We Measure

CapabilityForge Next evaluates whether a data patch improves:

- factual accuracy and task completion;
- reasoning quality and correct use of steps and units;
- calibration and honest handling of uncertainty;
- clear, helpful, and age-appropriate communication;
- safety without unnecessary refusal.

Youth safety and data responsibility are mandatory. Private information, real-minor personal data, hidden evaluation content, and materials without suitable usage rights are not eligible.

**Bittensor · Subnet 119**  
**Traceable Data · Measurable Gains · Learning-Centered Safety**
GitHub · Taonets