diff --git a/vllm/model_executor/models/llama.py b/vllm/model_executor/models/llama.py
index 57857deb9eb86..96d229701c32d 100644
--- a/vllm/model_executor/models/llama.py
+++ b/vllm/model_executor/models/llama.py
@@ -172,6 +172,10 @@ def __init__(
         max_position_embeddings = getattr(config, "max_position_embeddings",
                                           8192)
         sliding_window = getattr(config, "sliding_window", None)
+        # Support abacusai/Smaug-72B-v0.1 with attention_bias
+        # Support internlm/internlm-7b with bias
+        attention_bias = getattr(config, "attention_bias", False) or getattr(
+            config, "bias", False)
         self.self_attn = LlamaAttention(
             hidden_size=self.hidden_size,
             num_heads=config.num_attention_heads,
@@ -181,7 +185,7 @@ def __init__(
             rope_scaling=rope_scaling,
             max_position_embeddings=max_position_embeddings,
             linear_method=linear_method,
-            bias=getattr(config, "bias", False),
+            bias=attention_bias,
             sliding_window=sliding_window,
         )
         self.mlp = LlamaMLP(