card: results as charts (Jev 1.13, Kev-4B, Laya), 4K share card

README.md

@@ -20,7 +20,7 @@ An English decision model with the Jev API. You send a state and typed questions
 
 <video controls playsinline preload="none" poster="https://huggingface.co/datasets/mertkayacs/emberwick-videos/resolve/main/film/jevalt-film-en.jpg" src="https://huggingface.co/datasets/mertkayacs/emberwick-videos/resolve/main/film/jevalt-film-en-1080p.mp4"></video>
 
-*The one-minute film, sound on: three mistakes small decision models make and how JevAlt fixes each one.*
+*The 53-second film, sound on: two mistakes small decision models make and how JevAlt fixes each one.*
 
 ## Try it
 
@@ -68,12 +68,12 @@ client = TypeSafeClient(api_key="local", base_url="http://127.0.0.1:8000")
 
 ![Hidden instructions, option order, long policies and negated questions: JevAlt against Intern-Decision-4B, Kev-4B and Laya](https://huggingface.co/mertkayacs/Deem-4B/resolve/main/assets/fixes.png)
 
-Same items and client for every model, each as shipped: [Kev-4B](https://huggingface.co/jaredpalmer/kev-4b) r10 and [Laya](https://huggingface.co/convaiinnovations/laya) 0.3.22 on their own servers with their own calibration. Jev 1.13 rows come from [TypeSafe's notes](https://docs.typesafe.ai/model-jaggedness/jev-1.13) and an [independent audit](https://github.com/jujumilk3/jev-calibration-audit/blob/main/FINDINGS.md). The held-out tests come from JevAlt's own data pipeline, so they favour JevAlt. Kev-4B and Laya both do better on long padding; Laya is far smaller and faster. Every number and every decision: [results/comparison](https://huggingface.co/datasets/mertkayacs/jevalt-bench/tree/main/results/comparison).
+Same items and client for every model, each as shipped: [Kev-4B](https://huggingface.co/jaredpalmer/kev-4b) r10 and [Laya](https://huggingface.co/convaiinnovations/laya) 0.3.22 on their own servers with their own calibration. Jev 1.13 rows come from TypeSafe's [API reference](https://docs.typesafe.ai/api) and [Models page](https://docs.typesafe.ai/models). The held-out tests come from JevAlt's own data pipeline, so they favour JevAlt. Kev-4B and Laya both do better on long padding; Laya is far smaller and faster. Every number and every decision: [results/comparison](https://huggingface.co/datasets/mertkayacs/jevalt-bench/tree/main/results/comparison).
 
 <details>
 <summary><b>Significance and caveats</b></summary>
 
-![What Jev 1.13 lacks and JevAlt has: thinking when unsure, an unknown answer, coverage sets, native Turkish and German, open weights, repeatable answers](https://huggingface.co/mertkayacs/Deem-4B/resolve/main/assets/jev.png)
+![What Jev 1.13 lacks and JevAlt has: thinking when unsure, coverage sets, models made for Turkish and German, open weights](https://huggingface.co/mertkayacs/Deem-4B/resolve/main/assets/jev.png)
 
 The three test splits went through the same pipeline as the training rows, so they measure what the training aimed at. On the English split Deem-4B gains 4.4 accuracy points (paired bootstrap, 95% interval +3.4 to +5.5) and lowers Brier by 0.075; the Turkish and German splits move by +4.8 and +12.1 points. On JevBench-hard, TurkishMMLU and GermEval, which the training never saw, and on the typed-decisions test split (its train split was in the mix), accuracy does not change significantly, and Brier gets slightly worse on typed-decisions (+0.013) and 10kGNAD (+0.047). With `reasoning: "auto"` the English date, number and policy test rows go from 0.761 to 0.769 accuracy, an interval that touches zero.