Parameter	Value
start_layer_index	16
end_layer_index	28
preserve_good_behavior_weight	0.3326
steer_bad_behavior_weight	0.0048
overcorrect_relative_weight	1.0004
neighbor_count	15

Performance

Metric	This model	Original model (llm-jp/llm-jp-4-8b-thinking)
KL divergence	0.0129	0 (by definition)
Refusals	5/100	100/100

llm-jp-4-8b-thinking

LLM-jp-4 is a series of large language models developed by the Research and Development Center for Large Language Models at the National Institute of Informatics. This repository provides the llm-jp-4-8b-thinking model. For an overview of the LLM-jp-4 models across different parameter sizes, please refer to:

LLM-jp-4 Models Base models are trained with pre-training and mid-training only. Post-trained models are aligned using supervised fine-tuning (SFT) and direct preference optimization (DPO), without reinforcement learning. For practical usage examples and detailed instructions on how to use the models, please also refer to our cookbook.

Usage

python

import torch
from transformers import AutoModelForCausalLM, AutoTokenizer
model_name = "llm-jp/llm-jp-4-8b-thinking"
tokenizer = AutoTokenizer.from_pretrained(
    model_name,
    # trust_remote_code is required to load custom tokenizer and reasoning parser.
    trust_remote_code=True,
)
model = AutoModelForCausalLM.from_pretrained(
    model_name,
    dtype=torch.bfloat16,
    device_map="auto",
    trust_remote_code=True,
)
model.eval()
messages = [
    {"role": "user", "content": "自然言語処理とは何か"},
]
prompt: str = tokenizer.apply_chat_template(
    messages,
    tokenize=False,
    add_generation_prompt=True,
    reasoning_effort="medium",  # {"low", "medium", "high"}
)
inputs = tokenizer(prompt, return_tensors="pt").to(model.device)
with torch.no_grad():
    output_tensor = model.generate(
        **inputs,
        max_new_tokens=256,
        do_sample=True,
        temperature=0.7,
        top_p=0.9,
    )
generated_ids: list[int] = output_tensor[0][inputs["input_ids"].shape[1]:].tolist()
response = tokenizer.decode(generated_ids)
parsed = tokenizer.parse_response(response)
print("\n--- Parsed Response ---")
print("Role:", parsed.get("role"))
print("Thinking:", parsed.get("thinking"))
print("Content:", parsed.get("content"))

Model Details

Model type: Transformer-based Language Model
Architectures: Dense model: |Params|Layers|Hidden size|Heads|Context length|Embedding parameters|Non-embedding parameters|Total parameters| |:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:| |8B|32|4,096|32|65,536|805,306,368|7,784,894,464|8,590,200,832| MoE model: |Params|Layers|Hidden size|Heads|Routed Experts|Activated Experts|Context length|Embedding parameters|Non-embedding parameters|Activated parameters|Total parameters| |:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:|:---:| |32B-A3B|32|2,560|40|128|8|65,536|503,316,480|31,635,712,512|3,827,476,992|32,139,028,992|

Tokenizer

The tokenizer of this model is based on huggingface/tokenizers Unigram byte-fallback model. The vocabulary entries were converted from llm-jp-tokenizer v4.0. Please refer to README.md of llm-jp-tokenizer for details on the vocabulary construction procedure (the pure SentencePiece training does not reproduce our vocabulary).

[!NOTE] The chat template of this model is designed to be compatible with the OpenAI Harmony response format. However, the tokenizer differs from the one assumed by the openai-harmony library, and therefore direct tokenization with openai-harmony is not supported. For correct behavior, please use the tokenizer provided with this model. For detailed usage, please refer to our cookbook.

Training

Pre-training

This model is trained through a multi-stage pipeline consisting of pre-training and mid-training phases, using a total of 11.7T tokens.

The corpora used for pre-training and mid-training are publicly available at the following links:

[!NOTE] Although most of the corpora have been released, some portions are excluded from public release due to licensing constraints.

Post-training

We have fine-tuned the pre-trained checkpoint using SFT and further aligned it with DPO. The datasets used for post-training are also publicly available at the following links:

Evaluation

llm-jp-judge

We evaluated the model on a variety of tasks using an LLM-as-a-Judge framework. The descriptions of each task are as follows.

MT-Bench (JA/EN): A benchmark for measuring multi-turn conversational task-solving ability.
AnswerCarefully: A benchmark for evaluating safety in Japanese. We used 336 questions from the v2.0 test set.
llm-jp-instructions: A set of human-created single-turn question–answer pairs. We used 400 questions from the test set. We evaluated the models using gpt-5.4-2026-03-05.

[!NOTE] Note: In earlier evaluations of the llm-jp-3 series, we used gpt-4o-2024-08-06. The newer evaluator gpt-5.4-2026-03-05 provides a stricter and more reliable assessment, which results in lower scores on benchmarks such as MT-Bench compared to those reported for the llm-jp-3 series. The scores represent the average values obtained from three rounds of inference and evaluation. For more details, please refer to the codes. | Model Name | MT-Bench (JA) | MT-Bench (EN) | AnswerCarefully | llm-jp-instructions | |:-------------------------------------------------------------------------------------------------------|----:|----:|----------------:|--------------------:| | gpt-4o-2024-08-06 | 7.29 | 7.69 | 4.00 | 4.07 | | gpt-5.4-2026-03-05 (reasoning_effort = low) | 8.87 | 8.76 | 4.38 | 4.79 | | gpt-5.4-2026-03-05 (reasoning_effort = medium) | 8.87 | 8.89 | 4.43 | 4.82 | | gpt-5.4-2026-03-05 (reasoning_effort = high) | 8.98 | 8.85 | 4.41 | 4.83 | | gpt-oss-20b (reasoning_effort = low) | 7.21 | 7.95 | 3.39 | 3.08 | | gpt-oss-20b (reasoning_effort = medium) | 7.33 | 7.85 | 3.55 | 3.16 | | llm-jp-4-8b-thinking (reasoning_effort = low) | 7.23 | 7.54 | 3.58 | 3.50 | | llm-jp-4-8b-thinking (reasoning_effort = medium) | 7.54 | 7.79 | 3.69 | 3.54 | | llm-jp-4-32b-a3b-thinking (reasoning_effort = low) | 7.57 | 7.70 | 3.61 | 3.61 | | llm-jp-4-32b-a3b-thinking (reasoning_effort = medium) | 7.82 | 7.86 | 3.70 | 3.61 |

Risks and Limitations

The models released here are in the early stages of our research and development and have not been tuned to ensure outputs align with human intent and safety considerations.

Send Questions to

llm-jp(at)nii.ac.jp

License

Apache License, Version 2.0

Acknowledgement

To develop this model, we used the NINJAL Web Japanese Corpus (whole-NWJC) from the National Institute for Japanese Language and Linguistics (NINJAL).

Model Card Authors

The names are listed in alphabetical order. Hirokazu Kiyomaru and Takashi Kodama.

RumiaChannel/llm-jp-4-8b-thinking-uncensored-ara-gguf

Abliteration parameters