J

Japanese Novel Reward Modernbert Ja 130m

Developed by Aratako
A fine-tuned reward model based on modernbert-ja-130m for Japanese novel quality assessment, used to predict user ratings for novel texts.
Downloads 14
Release Time : 2/25/2025

Model Overview

This model predicts user ratings for input Japanese novel texts through regression, indirectly assessing text quality, primarily for applications like reinforcement learning in novel generation models.

Model Features

Long Text Processing Capability
Supports input texts up to 8192 tokens, suitable for processing complete novel chapters.
Quality Assessment
Indirectly evaluates novel text quality by predicting user ratings, typically outputting scores in the range of 0-10.
Reinforcement Learning Support
Designed specifically for reinforcement learning training scenarios in novel generation models, can be used as a reward signal.

Model Capabilities

Japanese Text Quality Assessment
Novel Rating Prediction
Long Text Processing

Use Cases

Text Generation
Novel Generation Model Training
Used as a reward model for training Japanese novel generation AI, optimizing generation quality through rating feedback.
Content Quality Screening
Automatically assesses the quality of user-generated novel content, aiding in content moderation.
Featured Recommended AI Models
ยฉ 2025AIbase