Research program
Small language models for Philippine languages
A technical research program investigating training and evaluation approaches for smaller language models designed around selected Philippine languages.
Questions in scope
- Which language communities and tasks can be studied responsibly with available evidence?
- How should data provenance, consent, and representativeness be documented?
- When can a smaller specialist model outperform a general model on a clearly bounded task?
Intended outputs
- Technical research note
- Model cards
- Evaluation artifacts
This work is in scoping. No model is currently being presented as an expert in any language, dialect, community, or domain.
Any future release will document the languages and tasks actually evaluated, training-data rights and provenance, known limitations, safety testing, and the conditions under which the model should not be used.