Unverified50% confidenceFactExact time
The base model is Qwen3-14B, from which two versions—KGModel and CGModel—are trained separately via supervised fine-tuning (SFT).
1
Sources
50%
Confidence
Medium-term (~90 days)
Relevance
9/14/2026
First Seen
Valid until: 12/13/2026
Sources
Related Entities
Related Claims
UnverifiedMainstream LLM products such as ChatGPT, DeepSeek, Kimi, and Qwen have all gone through three core stages: pre-training, SFT, and preference alignment.73% similarUnverifiedQwen2.5-27B是阿里云通义实验室发布的270亿参数开源模型，在MMLU、HumanEval、GSM8K等基准上接近GPT-4早期版本水平71% similarUnverified不同模型系列使用各自独立训练的词表，GPT系列采用标准BPE算法，Qwen系列针对中文与代码做了深度定制优化69% similarUnverifiedTEFM was tested on multiple mainstream large language models including Qwen3, Gemma-2, and Phi-4 to verify its model-agnostic nature.68% similarUnverifiedThe author open-sourced the fine-tuned Qwen3.5-4B Browser Agent SFT FP8 model, the Browser Agent framework, and two benchmarks: BrowseWebApp Bench and BU Bench V1.67% similar
Cite This Claim
Stable URI
https://kongchang.com/claim/919815API
curl https://kongchang.com/api/v1/knowledge/claims/919815MCP
get_claim(id=919815)