Datasets:
| # Using the FSOT Physical Archive for PFLT | |
| **Master:** `I:\FSOT-Physical-Archive\02_FSOT-2.1-Lean-Full` (I: definitive; GitHub is sync-from-I) | |
| ## Always | |
| ```text | |
| Law: vendor/fsot_compute.py pin D1D38A | |
| Scalar: S = K*(T1+T2+T3) from seeds only | |
| Domain: linguistic D_eff=12, observed=true for converse | |
| Students: densify/morph/phrase — never rewrite law | |
| ``` | |
| ## Grab for translation fluency | |
| | Archive path | Use | | |
| |--------------|-----| | |
| | `vendor/fsot_compute.py` | Authority pin + compute_scalar | | |
| | `vendor/linguistics/linguistics_derivations.json` | Zipf, dep length, sentence length | | |
| | `vendor/linguistics/data/LINGUISTIC_TARGETS.csv` | Empirical gates | | |
| | `data/linguistics_formal_benchmark.json` | D_eff=12 panel culture | | |
| | `FSOT/Scalar.lean` | Formal T1/T2/T3 structure | | |
| | `docs/PRACTICAL_PIPELINE.md` | Offline validation → application | | |
| ## Commands | |
| ```powershell | |
| cd pflt-Ada | |
| python fsot_archive_fluency_push.py | |
| python fsot_solve_fluency_gap.py | |
| ``` | |