Text Classification
Transformers
Safetensors
English
modernbert
ai-text-detection
idea-provenance
text-embeddings-inference
Instructions to use rishanthrajendhran/IdeaLens-ModernBERT-L-RolesOnly with libraries, inference providers, notebooks, and local apps. Follow these links to get started.
- Libraries
- Transformers
How to use rishanthrajendhran/IdeaLens-ModernBERT-L-RolesOnly with Transformers:
# Use a pipeline as a high-level helper from transformers import pipeline pipe = pipeline("text-classification", model="rishanthrajendhran/IdeaLens-ModernBERT-L-RolesOnly")# pip install -U transformers accelerate # Load model directly from transformers import AutoTokenizer, AutoModelForSequenceClassification tokenizer = AutoTokenizer.from_pretrained("rishanthrajendhran/IdeaLens-ModernBERT-L-RolesOnly") model = AutoModelForSequenceClassification.from_pretrained("rishanthrajendhran/IdeaLens-ModernBERT-L-RolesOnly", device_map="auto") - Notebooks
- Google Colab
- Kaggle
model card
Browse files
README.md
ADDED
|
@@ -0,0 +1,24 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
---
|
| 2 |
+
license: apache-2.0
|
| 3 |
+
tags: [ai-detection, text-classification]
|
| 4 |
+
extra_gated_prompt: "Access is granted individually. Please say who you are and what you intend to use the weights for."
|
| 5 |
+
---
|
| 6 |
+
|
| 7 |
+
# ideadet-modernbert-1m-roles
|
| 8 |
+
|
| 9 |
+
Idea-level AI-text detector. Answers **whose ideas a document contains**, not who typed the
|
| 10 |
+
sentences: a document whose ideas are a person's is labelled human however much of the surface
|
| 11 |
+
an AI produced.
|
| 12 |
+
|
| 13 |
+
- **Backbone:** modernbert
|
| 14 |
+
- **Training corpus:** 1m
|
| 15 |
+
- **Reads at inference:** the outline's role labels only, with the content removed
|
| 16 |
+
|
| 17 |
+
Stored `y=1` means HUMAN. The score is `softmax(logits)[:, 0]` = P(human), and the detector
|
| 18 |
+
fires when that falls below a calibrated threshold. Thresholds are quantiles of the score
|
| 19 |
+
distribution over held-out human documents and are **not** included here; a cut from one
|
| 20 |
+
model or input form is meaningless against another's scores.
|
| 21 |
+
|
| 22 |
+
Weights are float32, exactly as used to produce the reported numbers.
|
| 23 |
+
|
| 24 |
+
Fuller documentation and evaluation results to follow.
|