How to use from the
Use from the
Transformers library
# Use a pipeline as a high-level helper
from transformers import pipeline

pipe = pipeline("text-classification", model="PelaAI/KnowLine-4B-Gen5")
# pip install -U transformers accelerate
# Load model directly
from transformers import AutoProcessor, AutoModelForMultimodalLM

processor = AutoProcessor.from_pretrained("PelaAI/KnowLine-4B-Gen5")
model = AutoModelForMultimodalLM.from_pretrained("PelaAI/KnowLine-4B-Gen5", device_map="auto")
Quick Links

KnowLine-4B-Gen5

The fifth release of PelaAI's 4B System One decision model.

Inference guide ยท ไธญๆ–‡่ฏดๆ˜Ž ยท weights Apache-2.0 ยท previous version: KnowLine-4B-Gen4

Highlights:

  • Decision Index 0.3 public suite: 65.67 (self-run; Gen4: 64.90). The board's full score comes from the maintainers' private tests; see Comparison.
  • Calibrated probabilities: a temperature is folded into the weights. Computed with the board's method, ECE is about 0.04-0.05 and the Brier score about 0.28-0.31; see Calibration.
  • Fast: one request at a time on one H20, 31.6 ms median (see INFERENCE.md).
  • Images without any image training: the vision tower is the base model's and image input is zero-shot.

KnowLine is an independent model. It is not affiliated with, endorsed by, or derived from TypeSafe or Jev.

What it does

You send a state (text or a chat) and up to 64 typed questions: yes/no, choose one of k, or score on a rubric. The model answers every question with a probability distribution over its options:

  • one forward pass per question, with no text generation;
  • the output is the probability of each option's label token;
  • existing Jev clients only need to change the base URL, and users of earlier versions only need to change the model name, because the interface and serving settings are unchanged.
Base model Qwen/Qwen3.5-4B (Apache-2.0)
Training LoRA SFT (rank 32, alpha 64, language model only), merged into bf16 weights (see How Gen5 was made)
Release format bf16 weights, with a readout temperature T = 1.50 folded into the final norm weight (see Calibration). The vision tower and MTP head are unchanged from the base model; the config, tokenizer and chat template are identical to those of Gen1-Gen4.
Languages English, Simplified Chinese, Traditional Chinese

Quickstart

pip install "sglang==0.5.21" "transformers==5.12.1" requests
bash serve_knowline.sh PelaAI/KnowLine-4B-Gen5 0 8080     # SGLang (FP8 at load) on :9080 + /v1/systemone on :8080
curl -s http://127.0.0.1:8080/v1/systemone -H 'Content-Type: application/json' -d '{
  "model": "m",
  "state": "Customer: my order arrived broken, I want my money back.",
  "questions": {"refund": {"type": "noul", "instructions": "Should the agent offer a refund?"},
                "tone": {"type": "choice", "instructions": "Customer tone?",
                         "criteria": {"angry": "Angry", "neutral": "Neutral", "happy": "Happy"}}}}'
  • Front end: knowline_server.py is a single file that depends only on transformers and requests. It is the same file as in Gen4.
  • Full settings: the exact settings used for our Decision Index run are in INFERENCE.md.

How Gen5 was made

  • Base model: Qwen/Qwen3.5-4B.
  • Training: one epoch of LoRA SFT (rank 32, alpha 64, language model only), merged into bf16 weights. Training examples include decision tasks written and labelled by LLM teachers.
  • Readout temperature: T = 1.50, fitted on held-out data and folded into the final norm weight. It changes the size of the probabilities, but not which option ranks first.

The training data is not included in this repository.

Comparison

Scores on the Decision Index. Since Decision Index 0.4 (2026-10-19), the board's full score is 60% private tests of the same skills and 40% private tasks from new domains; the public suite is shown but no longer counts. Our full score is not available yet.

model size DI 0.3 public DI 0.4 full source
KnowLine-4B-Gen5 4B 65.67 not yet scored self-run, official kit
KnowLine-4B-Gen4 4B 64.90 not yet scored self-run, official kit
StartLux-Decision v1.1 4B 4.7B 67.81 56.35 board
Perplexity Decider Flash v1.2 4.5B 60.06 55.39 board
vLLM-SR d3-mini 4.5B 54.90 51.25 board
OneJev 1.6 4.5B 61.95 49.54 board
Jev 1.14 (API) 60.70 62.22 board
KnowLine-35B-A3B-Gen1 35B-A3B 69.49 62.28 board

Board rows as of the board's latest update (Decision Index 0.4).

Public-suite scores by area (DI 0.3):

area KnowLine-4B-Gen5 KnowLine-4B-Gen4
Knowledge & reasoning 45.0 44.6
Language 66.8 65.5
Retrieval & routing 73.3 72.5
Tools & agents 87.4 87.6
Arts & taste 61.2 59.2

Releases

The model is trained in a self-evolving loop, and new versions will follow.

model date DI 0.3 public notes
KnowLine-4B-Gen5 (this model) 2026-10-21 65.67 fifth release
KnowLine-4B-Gen4 2026-10-10 64.90 fourth release
KnowLine-4B-Gen3 2026-10-09 63.11 third release
KnowLine-4B-Gen2 2026-10-08 62.54 second release
KnowLine-4B-Gen1 2026-10-07 60.47 first release

Evaluation

All results are self-run and have not been verified by a third party.

Other public benchmarks

suite Gen5 Gen4
C-Eval (4 categories, macro) 79.2 78.7
JevBench / JevBench-hard 83.1 / 66.7 86.1 / 72.1

Web operation (Mind2Web official test splits, evaluation only)

split Gen5 element selection Gen5 operation (balanced) Gen4 element selection Gen4 operation (balanced)
test_task (Cross-Task) 92.6 97.1 92.9 97.0
test_website (Cross-Website) 90.9 96.9 91.9 95.7
test_domain (Cross-Domain) 91.8 98.4 91.7 97.2
  • About the same as Gen4; the differences are within noise.
  • The task is to pick the target element from a candidate set: the target plus up to 5 other elements sampled from the page. This is easier than the original Mind2Web protocol, so do not compare these results with the Mind2Web leaderboard.

Images (zero-shot)

  • No image training: Gen5 keeps the base model's vision tower unchanged and saw no images in training. knowline_server.py accepts OpenAI-style image / image_url content parts in a chat state and passes them to SGLang; each question is still one prefill.

Calibration

Computed on the 0.3 public suite using the method described by the Decision Index board: each field is scored right or wrong, confidence is the probability of the chosen option, benchmarks are weighted equally, and 10 equal-width bins are used.

metric Gen5 Gen4
ECE (lower is better) 0.04-0.05 0.04-0.05
Brier score (lower is better) 0.28-0.31 0.29-0.31
Confidence โ‰ฅ95% but wrong 1.5-1.6% 1.6-1.8%
Mean confidence / accuracy 0.82-0.83 / 0.78-0.80 0.82 / 0.77
  • Temperature in the weights: the final norm weight includes a readout temperature T = 1.50, so every logit is divided by 1.50 before the softmax over the option labels. The top-ranked option does not change; the probabilities are just less extreme. T was fitted on held-out data.
  • Why ranges: the board does not publish which 32 benchmarks it uses, so we report ranges over the plausible benchmark sets, as for the earlier releases.
  • Recommendation: if your use depends on probability thresholds, check calibration on your own data.

Disclosures

  • Decontamination: all training data was checked against every Decision Index 0.3 row and the public benchmarks we report (Intern-Decision, S1MB, JevBench, MMDM). About 22,000 rows were removed: rows that nearly duplicate a benchmark item, repeat its question or claim, or contain its whole state, and, in a stricter second pass, most rows that share a source passage with a benchmark item.
  • Teacher-labelled synthetic data: LLM teachers wrote and labelled our synthetic tasks. The labels were filtered, but not all of them were checked by a human.
  • Public suite: the Decision Index public suite is public; the board's private tests may show a lower score.

Limitations

  • Knowledge-heavy reasoning: weaker than larger models. The knowledge area score is 45.0 (DI 0.3); a 4B model is limited here, and the new-domain part of the board favours larger models.
  • JevBench-hard: 66.7 against 72.1 for Gen4; the drop is not significant, but check it if your use is close to JevBench-hard.
  • New domains: the board's private new-domain tasks favour larger models; check the model on your own domain.
  • Prompt injection: an instruction planted in the state can change the answer. Keep untrusted text clearly delimited.
  • Images: there was no image training; image use is zero-shot, so check it on your own data before relying on it.

Citation

@misc{knowline4bgen5,
  title  = {KnowLine-4B-Gen5: a 4B decision model},
  author = {PelaAI},
  year   = {2026},
  url    = {https://huggingface.co/PelaAI/KnowLine-4B-Gen5}
}

License

  • Weights: Apache-2.0, the same as the base model.
  • Code: knowline_server.py is MIT (see the file header).
Downloads last month
281
Safetensors
Model size
5B params
Tensor type
F32
ยท
BF16
ยท
Inference Providers NEW
This model isn't deployed by any Inference Provider. ๐Ÿ™‹ Ask for provider support

Model tree for PelaAI/KnowLine-4B-Gen5

Finetuned
Qwen/Qwen3.5-4B
Finetuned
(1,485)
this model
Quantizations
1 model