PsiQRH Riemann Sphere - Benchmark Results
收藏官方服务：
资源简介：
PsiQRH Riemann Sphere language model benchmark results. 32.1M params, Riemann sphere attention (96 params), GPT-2 BPE tokenizer. Compared against GPT-2 (124M) and random baseline across 5 domains (code, poetry, math, dialogue, science), latency profiling, and GSM8K reasoning.
提供机构：
Zenodo创建时间：
2026-06-18
