Same prompt, all 22 Turbo Brilliance reasoning modes, one small local model on one GPU: pick your reasoning effort with measured numbers, not vibes
-
Updated
Oct 31, 2026 - JavaScript
Same prompt, all 22 Turbo Brilliance reasoning modes, one small local model on one GPU: pick your reasoning effort with measured numbers, not vibes
Benchmark testing whether catch22 time-series features can detect genuine chaos even when linear statistics (mean, variance, spectrum) are deliberately matched via IAAFT surrogate decoys.
CSCE 5222 Feature Engineering - Group 18 Project 2: Hybrid, Automated, and Interpretable Feature Engineering for Time-Series
To associate your repository with the catch22 topic, visit your repo's landing page and select "manage topics."