Upload README.md with huggingface_hub

README.md

@@ -89,7 +89,7 @@ Results are **split by reasoning mode**: comparing a thinking-on (reasoning) mod
 
 ⁹ Six-rung GGUF quant-tax ladder (bartowski imatrix repo), measured 2026-08-12/13, llama.cpp b10338 worktree, all rungs fully VRAM-resident. The tax is nearly zero above the floor: Q5_K_M to IQ2_M spans 0.4 q_avg while decode climbs 349 → 399 tok/s and VRAM drops 25.2 → 17.7 GiB. Only IQ2_XXS pays (−1.7 q_avg vs IQ2_M), and the entire drop is HellaSwag. IQ2_M strictly dominates Q3_K_M and IQ4_XS (smaller, higher q_avg, faster). Ladder GGUFs embed the MTP drafter head (inert here). [Ladder report](reports/nemotron-lightning-quant-ladder.md).
 
-¹⁰ Seven-rung GGUF quant-tax ladder (unsloth day-0 repo, UD = dynamic imatrix recipe), measured 2026-08-14/15, llama.cpp b9653, all rungs fully VRAM-resident. No cliff anywhere: Q8_0 to UD-IQ2_XXS spans 2.9 q_avg over a 3.4x file-size range, Q6_K ties Q8_0 to the second decimal (93.66 both), and UD-IQ3_XXS holds 92.7 from an 11.1GB file (12.8 GiB peak — 16GB-card territory) at 96 tok/s. MMLU pays most of the tax; the IQ2 rungs pinch HumanEval first. [Ladder report](reports/qwen3-8-27b-quant-ladder.md).
+¹⁰ Seven-rung GGUF quant-tax ladder (unsloth day-0 repo, UD = dynamic imatrix recipe), measured 2026-08-14/15, llama.cpp b9653, all rungs fully VRAM-resident. No cliff anywhere: Q8_0 to UD-IQ2_XXS spans 2.9 q_avg over a 3.4x file-size range, Q6_K ties Q8_0 to the second decimal (93.66 both), and UD-IQ3_XXS holds 92.7 from an 11.1GB file (12.8 GiB peak — 16GB-card territory) at 96 tok/s. MMLU pays most of the tax; the IQ2 rungs pinch HumanEval first. The tightness is short-context only: on long-context retrieve-and-use (15 tasks per depth at 16k/32k/64k, measured 2026-08-16) Q6_K is perfect at every depth while UD-IQ3_XXS holds 73/67/80 — the 1.0 q_avg gap becomes ~27 points under depth load. [Ladder report + depth addendum](reports/qwen3-8-27b-quant-ladder.md).
 
 ### Thinking ON (reasoning · extended chain-of-thought)