autotrust-jev:27b-vl-nvfp4

308 4 hours ago

AutoTrust’s decision models return choices, yes/no probabilities, and ratings for the questions you supply

vision decision 9b 35b 
curl http://localhost:11434/v1/systemone \
  -H "Content-Type: application/json" \
  -d '{
    "model": "autotrust-jev:27b-vl-nvfp4",
    "state": "Hello World",
    "questions": {
      "says_hello": {
        "type": "noul",
        "instructions": "Does the state text contain a greeting?",
        "criteria": {
          "true": "The state text contains a greeting.",
          "false": "The state text does not contain a greeting."
        }
      }
    }
  }'

Details

4 hours ago

ddd4abaa2baf · 21GB

json
{ "version": 1, "per_kind": { "noul": 1.0143134751376188, "choice": 1.0161280671450397, "score": 1.0
json
{ "architectures": [ "AutoTrustDecision" ], "dtype": "bfloat16", "image_token_id": 248056, "language
json
{ "bias": [ 0.00011960109259234741, -0.00011981795250903815, -0.001826660125516355, -0.0008995002135
json
{ "bos_token_id": 248044, "do_sample": true, "eos_token_id": [ 248046, 248044 ], "pad_token_id": 248
json
{ "producer": { "name": "modelopt", "version": "0.47.0rc1.dev283+gf1d68fd4d" }, "quantization": { "q
json
{ "size": { "longest_edge": 16777216, "shortest_edge": 65536 }, "patch_size": 16, "temporal_patch_si
json
{ "version": "1.0", "truncation": null, "padding": null, "added_tokens": [ { "id": 248044, "content"
json
{ "add_prefix_space": false, "added_tokens_decoder": { "248044": { "content": "<|endoftext|>", "lstr
json
{ "size": { "longest_edge": 25165824, "shortest_edge": 4096 }, "patch_size": 16, "temporal_patch_siz
json
{ "!": 0, "\"": 1, "#": 2, "$": 3, "%": 4, "&": 5, "'": 6, "(": 7, ")": 8, "*": 9, "+": 10, ",": 11,
license
Apache License Version 2.0, January 2004 http://www.apache.org/licenses/ TERMS AND CONDITIONS FOR US
params
{ "temperature": 1, "top_k": 20, "top_p": 0.95 }
1199 tensors

Readme

AutoTrust’s decision models return choices, yes/no probabilities, and ratings for the questions you supply. Use them to route requests, classify documents, or grade an answer against your own criteria.

  • Decisions from your criteria: define the categories or six rating levels that matter to your workflow.
  • Probabilities for every option: use the full distribution when deciding what to automate or send for review.
  • Several questions, one input: ask about the same text or structured state in a single request.

Try it

Route a customer request to the right team:

ollama run autotrust-jev "My card was charged twice for the same order. Please refund the duplicate charge." \
  --ask "Which team should handle this request?" \
  --choice "billing=Charges, payments, and refunds" \
  --choice "shipping=Deliveries and tracking" \
  --choice "technical=App errors and account access"

autotrust-jev runs AutoTrust’s 9B text model. For decisions about images, choose autotrust-jev:27b-vl or autotrust-jev:35b.

See the decision documentation for CLI options and the API.

Publisher results

AutoTrust reports these results for its 9B checkpoint on its held-out test_set_30k evaluation:

Task Publisher result
Yes/no AUROC, all evaluation rows 0.996
Rating mean absolute error, 0–5 scale 0.103

See AutoTrust’s evaluation for the target distributions, data sources, and methodology.

Credits

Published by AutoTrust: 9B, 27B-VL, and 35B.

Released under Apache 2.0.