02 / The lab
What would count as evidence?
I use this space to make questions testable. Two open-source tools verify what can be discarded and which sources support an AI answer. Two studies ask whether model scores can signal uncertainty and which patterns in the Indus script survive a statistical baseline.
Open-source tools and studies / 04
- URP How much work can a system skip while preserving the result you need? Released v0.1.0
- RDLLM Which source supports each claim in an AI answer, and who should receive credit? Released v1.0.0 · Beta
- Kaprekar Geometry Could the gaps between a model’s scores flag an answer worth checking? Manuscript in preparation
- Indus Script Which sign patterns survive a statistical baseline in the Indus script? Unreviewed preprint