Upload folder using huggingface_hub

README.md

@@ -45,13 +45,13 @@ Profiles: `chat-medium`, `chat-multiturn-long`, `chat-multiturn-medium`, `chat-m
 
 ### distributional
 
-Requests sample ISL/OSL from parameterized statistical distributions (e.g., lognormal) fitted to real workload statistics. Provides controlled experimental variation while matching aggregate characteristics.
+Requests sample ISL/OSL from parameterized statistical distributions (e.g., lognormal) fitted to real workload statistics. Shorter to run than full trace replays, enabling faster characterization of profile-level serving behavior across the hardware matrix. MSE validation confirms distributional runs reproduce the latency and throughput characteristics of their trace-replay counterparts.
 
 Profiles: `chat-multiturn`, `chat-singleturn`, `coding-singleturn`, `osworld-multiturn`, `swebench-multiturn`, `terminalbench-multiturn`
 
 ### Why two configurations?
 
-**trace_replay** proves relevance: patterns are drawn from real agent sessions, grounding results in observed behavior. **distributional** proves mechanism: controlled statistical sampling isolates the effects of workload parameters from incidental correlations. Together they support both ecological validity and experimental control.
+**trace_replay** provides ecological validity — patterns are drawn from real agent sessions, grounding results in observed behavior. **distributional** enables efficient coverage — shorter run times allow systematic sweeps across the full model-hardware-concurrency matrix, while MSE validation against trace_replay confirms the results remain representative.
 
 ### Concurrency filtering
 

croissant.json

@@ -225,7 +225,7 @@
       "@type": "cr:RecordSet",
       "@id": "distributional-summary",
       "name": "Distributional Summary",
-      "description": "One row per benchmark configuration from distributional runs (statistical sampling from fitted distributions).",
+      "description": "One row per benchmark configuration from distributional runs (statistical sampling, shorter to run, MSE-validated against trace replays).",
       "field": [
         {
           "@type": "cr:Field",