Update README.md

README.md

@@ -343,8 +343,8 @@ base_model = AutoModelForCausalLM.from_pretrained(
     torch_dtype=torch.bfloat16, 
     device_map="auto"
 )
-model = PeftModel.from_pretrained(base_model, "Phase-Technologies/ReasonBorn-Qwen-3B")
-tokenizer = AutoTokenizer.from_pretrained("Phase-Technologies/ReasonBorn-Qwen-3B")
+model = PeftModel.from_pretrained(base_model, "Xerv-AI/ReasonBorn-Qwen-3B")
+tokenizer = AutoTokenizer.from_pretrained("Xerv-AI/ReasonBorn-Qwen-3B")
 
 prompt = """<|im_start|>system
 You are ReasonBorn. Use <plan>, <reasoning> with <step> & <verify>, <conclusion> strictly.
@@ -369,6 +369,6 @@ print(tokenizer.decode(output[0], skip_special_tokens=True))
 
 ---
 
-**Phase-Technologies / ReasonBorn-Qwen-3B**  
+**Xerv-AI / ReasonBorn-Qwen-3B**  
 First step toward verifiable, tagged, auditable AI mathematical reasoning.  
 Trained in Kolkata, March 2026.