cloudyu commited on
Commit
7cbcb19
·
verified ·
1 Parent(s): 0009baa

Model card: Decision Index #1 first, then the model introduction

Browse files
Files changed (1) hide show
  1. README.md +10 -10
README.md CHANGED
@@ -24,16 +24,6 @@ tags:
24
 
25
  # autotrust/JEV-27B-VL
26
 
27
- ### JEV-27B that can see: the same System 1 decisions and System 2, now on images
28
-
29
- **autotrust/JEV-27B-VL** is [autotrust/JEV-27B](https://huggingface.co/autotrust/JEV-27B) with vision, and the
30
- **#1 model on the [Jev Decision Index Vision board](https://huggingface.co/spaces/multimodalart/jev-decision-index?mode=vision)**.
31
-
32
- | | what it does | output |
33
- |---|---|---|
34
- | **System 1** (`POST /v1/decide`) | typed decisions: yes/no · pick one of 2–256 options · rate 0–5, over text and images; prompts up to 256K tokens | a calibrated probability for every option, in one forward pass |
35
- | **System 2** (`autotrust/JEV-27B-VL`) | the unmodified Qwen3.8-27B, optionally thinking step by step, with image input | text / reasoning |
36
-
37
  ## #1 on the Decision Index Vision board
38
 
39
  JEV-27B-VL is **ranked first of 20 vision decision models** on the community
@@ -90,6 +80,16 @@ same skills.
90
  **In one sentence:** the most accurate of 20 vision decision models, with no weak skill group, first on unseen images,
91
  calibrated, fast, and above a 397B model at a fifteenth of the size, from a decision head trained on text alone.
92
 
 
 
 
 
 
 
 
 
 
 
93
  ## New (3 October 2026): robot arm and computer use
94
 
95
  Every step below is **one System 1 decision**: a camera image or a screenshot in, a probability for every action out, in
 
24
 
25
  # autotrust/JEV-27B-VL
26
 
 
 
 
 
 
 
 
 
 
 
27
  ## #1 on the Decision Index Vision board
28
 
29
  JEV-27B-VL is **ranked first of 20 vision decision models** on the community
 
80
  **In one sentence:** the most accurate of 20 vision decision models, with no weak skill group, first on unseen images,
81
  calibrated, fast, and above a 397B model at a fifteenth of the size, from a decision head trained on text alone.
82
 
83
+ ## JEV-27B that can see: the same System 1 decisions and System 2, now on images
84
+
85
+ **autotrust/JEV-27B-VL** is [autotrust/JEV-27B](https://huggingface.co/autotrust/JEV-27B) with vision, and the
86
+ **#1 model on the [Jev Decision Index Vision board](https://huggingface.co/spaces/multimodalart/jev-decision-index?mode=vision)**.
87
+
88
+ | | what it does | output |
89
+ |---|---|---|
90
+ | **System 1** (`POST /v1/decide`) | typed decisions: yes/no · pick one of 2–256 options · rate 0–5, over text and images; prompts up to 256K tokens | a calibrated probability for every option, in one forward pass |
91
+ | **System 2** (`autotrust/JEV-27B-VL`) | the unmodified Qwen3.8-27B, optionally thinking step by step, with image input | text / reasoning |
92
+
93
  ## New (3 October 2026): robot arm and computer use
94
 
95
  Every step below is **one System 1 decision**: a camera image or a screenshot in, a probability for every action out, in