GLM-5.2 is live. #1 throughput on OpenRouter, pay-per-token on FriendliAI. Try it today ➜
hotdogs
Available on FriendliAI
Run this model inference on single tenant GPU with unmatched speed and reliability at scale.
Model Details
Model Provider
hotdogs
Model Tree
InternScience/Agents-A1-4B
Input Modalities
Output Modalities
Supported Functionality
| Capability | Description |
|---|---|
| 🛡️ Cyber Expert | Offensive + defensive security, penetration testing, red/blue team |
| 🧠 Causal Analysis | Step-by-step security reasoning (MITRE ATT&CK, OWASP) |
| 🚨 Incident Response | Detection, containment, and response playbooks |
| 🔍 Threat Intel | CVE analysis, IOC correlation, ransomware indicators |
| 💻 Code & Exploit | Python/Bash security tooling, exploit analysis |
| 🎛️ Adjustable | --lora-scaled lets you control LoRA strength (0.1–0.7+) |
| 🌏 Thai + English | Native bilingual base model support |
| Parameter | Value |
|---|---|
| Base Model | InternScience/Agents-A1-4B |
| Parameters | ~4.62B (4.29B trainable-excluded) |
| Architecture | Qwen3.5 hybrid (Linear + Full attention) |
| LoRA Rank | r=64, alpha=128, dropout=0.0 |
| Target Modules | q/k/v/o_proj, gate/up/down_proj (all linear) |
| Training Data | 168,793 samples (Fenrir + Trendyol + ShareGPT) |
| Max Seq Length | 4096 tokens |
| Training Steps |
Trained on 3 curated cybersecurity datasets (all English, unified system prompt):
| Dataset | Rows | Content |
|---|---|---|
| AlicanKiraz0/Cybersecurity-Dataset-Fenrir-v2.1 | 99,869 | Causal analysis, threat intel, MITRE, ransomware, red/blue team |
| Trendyol/Trendyol-Cybersecurity-Instruction-Tuning-Dataset | 53,201 | Cyber-defense, C2 analysis, DLP, honeypot, insider threat |
| ChaoticNeutrals/Cybersecurity-ShareGPT | 15,723 | Cybersecurity Q&A + step-by-step reasoning |
Training details:
This repo ships gguf/cyber_lora_v1.gguf — a GGUF-format LoRA adapter that works with any
Agents-A1-4B base GGUF via --lora-scaled.
Tune the strength with the scale value:
0.1 — subtle cyber knowledge boost (keeps base personality)0.3 — balanced security-aware assistant0.5 — strong cybersecurity expert behavior0.7 — heavy security specialization (may override base style)>0.7 — possible overfit artifacts / style override; test per use casebash
# With our Fable heretic GGUF (recommended — uncensored + cyber)llama-server \-m /models/Agents-A1-4B-Fable-Preview-heretic-F16.gguf \--lora-scaled /models/cyber_lora_v1.gguf:0.5 \--host 0.0.0.0 --port 8080 \--n-gpu-layers 999 \--ctx-size 32768 \--flash-attn on \--cont-batching \--mlock \--temp 0.95 \--top-k 40 \--top-p 0.9 \--min-p 0.0 \-n -1 \--no-mmap \--parallel 1 --tools all \--dry-multiplier 0.0 \--jinja --dry-sequence-breaker none --repeat-penalty 1.1# Or with our Kimi heretic GGUF (coding-focused base)llama-server \-m /models/Agents-A1-4B-kimi-Preview-heretic-IQ4_NL.gguf \--lora-scaled /models/cyber_lora_v1.gguf:0.5 \--host 0.0.0.0 --port 8080 \--n-gpu-layers 999 \--ctx-size 32768 \--flash-attn on \--cont-batching \--mlock \--temp 0.95 \--top-k 40 \--top-p 0.9 \--min-p 0.0 \-n -1 \--no-mmap \--parallel 1 --tools all \--dry-multiplier 0.0 \--jinja --dry-sequence-breaker none --repeat-penalty 1.1# CLI quick testllama-cli -m Agents-A1-4B-Fable-Preview-heretic-F16.gguf \--lora-scaled cyber_lora_v1.gguf:0.5 \-p "What is SQL injection and how do I prevent it?" -n 256
💡 Compatible base models (any Agents-A1-4B GGUF works):
Since LoRA only adds weights, you can also apply it to the plain InternScience/Agents-A1-4B GGUF of your choice.
python
from transformers import AutoModelForCausalLM, AutoTokenizerfrom peft import PeftModelbase = "InternScience/Agents-A1-4B"adapter = "hotdogs/Agents-A1-4B-cyber-lora"model = AutoModelForCausalLM.from_pretrained(base, torch_dtype="auto", device_map="auto", trust_remote_code=True)model = PeftModel.from_pretrained(model, adapter)tokenizer = AutoTokenizer.from_pretrained(adapter)messages = [{"role": "system", "content": "You are an elite AI security expert specializing in offensive security, defensive strategies, threat intelligence, incident response, and vulnerability analysis."},{"role": "user", "content": "What is SQL injection and how do I prevent it?"},]inputs = tokenizer.apply_chat_template(messages, tokenize=True, return_tensors="pt")outputs = model.generate(inputs, max_new_tokens=256, temperature=0.6)print(tokenizer.decode(outputs[0]))
| Prompt (Thai) | English Equivalent |
|---|---|
SQL injection คืออะไร ป้องกันยังไง | What is SQL injection and how do I prevent it? |
ตรวจจับ lateral movement ด้วย Windows Event ID ยังไง | How would you detect lateral movement using Windows Event ID correlation? |
MITRE ATT&CK ใช้ยังไงใน incident response | How is MITRE ATT&CK used in incident response? |
สัญญาณ ransomware มีอะไรบ้าง SOC ต้องทำยังไง | What are common indicators of ransomware and how should a SOC respond? |
bash
# Thai prompt via llama-clillama-cli -m Agents-A1-4B-Fable-Preview-heretic-F16.gguf \--lora-scaled cyber_lora_v1.gguf:0.5 \--flash-attn on --cont-batching --mlock \--temp 0.95 --top-k 40 --top-p 0.9 --min-p 0.0 \--no-mmap --parallel 1 --tools all \--dry-multiplier 0.0 --jinja --dry-sequence-breaker none \-p "SQL injection คืออะไร ป้องกันยังไง" -n 256
| File | Size | Description |
|---|---|---|
adapter_model.safetensors | 340 MB | PEFT LoRA adapter (BF16) |
gguf/cyber_lora_v1.gguf | 339 MB | GGUF LoRA for llama.cpp --lora-scaled |
tokenizer.json | 20 MB | Tokenizer |
chat_template.jinja | 8.98 kB |
This model is fine-tuned on cybersecurity content including offensive techniques. Use it only for authorized security testing, education, and defensive purposes. The authors are not responsible for any misuse. Always follow applicable laws and obtain permission before testing systems you do not own.
If you find this model useful, please consider supporting my work! หากคุณคิดว่าโมเดลนี้มีประโยชน์ กรุณาสนับสนุนผลงานของฉันด้วยนะคะ! 🙏
markdown
bc1qf27cyk3vmugcdyv9xdtuv5jwz37863crpj5c9v
Thank you for your support! 🙏✨ ขอบคุณมากๆ สำหรับการสนับสนุนค่า! 💖🤗
Built with ❤️ by UKA — 18-year-old coder & cybersecurity expert
| 40,089 (1 epoch) |
| Format | ChatML (Jinja2 template) |
| ChatML Jinja template |