Text Classification
Transformers
Safetensors
English
Chinese
qwen3_5_moe_text
text-generation
decision-model
web-agent
browser-agent
typed-decisions
structured-output
one-pass
mixture-of-experts
Instructions to use Lexmount/WebJev-35B-A3B with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use Lexmount/WebJev-35B-A3B with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="Lexmount/WebJev-35B-A3B")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForCausalLM tokenizer = AutoTokenizer.from_pretrained("Lexmount/WebJev-35B-A3B") model = AutoModelForCausalLM.from_pretrained("Lexmount/WebJev-35B-A3B", device_map="auto") - Notebooks
- Google Colab
- Kaggle
Model card: link the WebJev dataset
Browse files
README.md
CHANGED
|
@@ -19,7 +19,7 @@ tags:
|
|
| 19 |
- one-pass
|
| 20 |
- mixture-of-experts
|
| 21 |
datasets:
|
| 22 |
-
- Lexmount/
|
| 23 |
---
|
| 24 |
|
| 25 |
# WebJev-35B-A3B
|
|
@@ -269,7 +269,7 @@ The training mixture covers:
|
|
| 269 |
|
| 270 |
- **General typed decisions.** Public classification, entailment, question-answering, routing, tool-use and
|
| 271 |
policy-rule tasks in the System One question types (choice, noul, score), following the open Decider recipe.
|
| 272 |
-
- **Web-agent decisions.** [
|
| 273 |
execution-verified action-prediction and element-grounding decisions on 1,443 live websites.
|
| 274 |
- **Open-Jev typed decisions.** Browser control, reasoning, entity alignment, extraction and customer-support tasks.
|
| 275 |
- **Knowledge and science.** Multiple-choice questions from Nemotron-RL-knowledge-mcqa and OpenScienceReasoning-2.
|
|
|
|
| 19 |
- one-pass
|
| 20 |
- mixture-of-experts
|
| 21 |
datasets:
|
| 22 |
+
- Lexmount/WebJev
|
| 23 |
---
|
| 24 |
|
| 25 |
# WebJev-35B-A3B
|
|
|
|
| 269 |
|
| 270 |
- **General typed decisions.** Public classification, entailment, question-answering, routing, tool-use and
|
| 271 |
policy-rule tasks in the System One question types (choice, noul, score), following the open Decider recipe.
|
| 272 |
+
- **Web-agent decisions.** [WebJev](https://huggingface.co/datasets/Lexmount/WebJev): 64K
|
| 273 |
execution-verified action-prediction and element-grounding decisions on 1,443 live websites.
|
| 274 |
- **Open-Jev typed decisions.** Browser control, reasoning, entity alignment, extraction and customer-support tasks.
|
| 275 |
- **Knowledge and science.** Multiple-choice questions from Nemotron-RL-knowledge-mcqa and OpenScienceReasoning-2.
|