diff --git a/src/alpaca_eval/models_configs/LMCocktail-10.7B-v1/configs.yaml b/src/alpaca_eval/models_configs/LMCocktail-10.7B-v1/configs.yaml
index 5b97a047..a0a60a67 100644
--- a/src/alpaca_eval/models_configs/LMCocktail-10.7B-v1/configs.yaml
+++ b/src/alpaca_eval/models_configs/LMCocktail-10.7B-v1/configs.yaml
@@ -9,7 +9,7 @@ LMCocktail-10.7B-v1:
     max_new_tokens: 2048
     temperature: 0.7
     top_p: 1.0
-    do_sample: True
+    use_beam_search: True
     length_penalty: 2.0
     earlth_stopping: True
   pretty_name: "LMCocktail-10.7B-v1"
diff --git a/src/alpaca_eval/models_configs/Llama-3-Instruct-8B-SimPO/configs.yaml b/src/alpaca_eval/models_configs/Llama-3-Instruct-8B-SimPO/configs.yaml
index 23a0c57a..057da030 100644
--- a/src/alpaca_eval/models_configs/Llama-3-Instruct-8B-SimPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/Llama-3-Instruct-8B-SimPO/configs.yaml
@@ -1,7 +1,7 @@
 Llama-3-Instruct-8B-SimPO:
   completions_kwargs:
     batch_size: 900
-    do_sample: true
+    use_beam_search: true
     max_new_tokens: 4096
     model_kwargs:
       dtype: bfloat16
diff --git a/src/alpaca_eval/models_configs/Starling-LM-7B-alpha-ExPO/configs.yaml b/src/alpaca_eval/models_configs/Starling-LM-7B-alpha-ExPO/configs.yaml
index 13798548..7d1e4fc1 100644
--- a/src/alpaca_eval/models_configs/Starling-LM-7B-alpha-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/Starling-LM-7B-alpha-ExPO/configs.yaml
@@ -7,7 +7,7 @@ Starling-LM-7B-alpha-ExPO:
       dtype: 'bfloat16'
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/Starling-LM-7B-beta-ExPO/configs.yaml b/src/alpaca_eval/models_configs/Starling-LM-7B-beta-ExPO/configs.yaml
index c949aca8..eb697a52 100644
--- a/src/alpaca_eval/models_configs/Starling-LM-7B-beta-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/Starling-LM-7B-beta-ExPO/configs.yaml
@@ -7,7 +7,7 @@ Starling-LM-7B-beta-ExPO:
       dtype: 'bfloat16'
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/causallm-14b/configs.yaml b/src/alpaca_eval/models_configs/causallm-14b/configs.yaml
index 10faa6ad..95cf1881 100644
--- a/src/alpaca_eval/models_configs/causallm-14b/configs.yaml
+++ b/src/alpaca_eval/models_configs/causallm-14b/configs.yaml
@@ -8,7 +8,7 @@ causallm-14b:
     max_new_tokens: 2048
     temperature: 0.7
     top_p: 1.0
-    do_sample: False
+    use_beam_search: False
     batch_size: 8
   pretty_name: "CausalLM-14B"
   link: "https://huggingface.co/CausalLM/14B"
diff --git a/src/alpaca_eval/models_configs/deita-7b-v1.0/configs.yaml b/src/alpaca_eval/models_configs/deita-7b-v1.0/configs.yaml
index da075009..2b0d9b00 100644
--- a/src/alpaca_eval/models_configs/deita-7b-v1.0/configs.yaml
+++ b/src/alpaca_eval/models_configs/deita-7b-v1.0/configs.yaml
@@ -8,6 +8,6 @@ deita-7b-v1.0:
     max_new_tokens: 2048
     temperature: 0.7
     top_p: 1.0
-    do_sample: True
+    use_beam_search: True
   pretty_name: "DEITA 7B v1.0"
   link: "https://github.com/hkust-nlp/deita"
\ No newline at end of file
diff --git a/src/alpaca_eval/models_configs/internlm2-chat-20b-ExPO/configs.yaml b/src/alpaca_eval/models_configs/internlm2-chat-20b-ExPO/configs.yaml
index bfb0135a..8d5102b6 100644
--- a/src/alpaca_eval/models_configs/internlm2-chat-20b-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/internlm2-chat-20b-ExPO/configs.yaml
@@ -8,7 +8,7 @@ internlm2-chat-20b-ExPO:
       trust_remote_code: True
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/internlm2-chat-7b-ExPO/configs.yaml b/src/alpaca_eval/models_configs/internlm2-chat-7b-ExPO/configs.yaml
index 294d87f6..580b0b29 100644
--- a/src/alpaca_eval/models_configs/internlm2-chat-7b-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/internlm2-chat-7b-ExPO/configs.yaml
@@ -8,7 +8,7 @@ internlm2-chat-7b-ExPO:
       trust_remote_code: True
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/merlinite-7B-AOT/configs.yaml b/src/alpaca_eval/models_configs/merlinite-7B-AOT/configs.yaml
index 44d68023..4098058e 100644
--- a/src/alpaca_eval/models_configs/merlinite-7B-AOT/configs.yaml
+++ b/src/alpaca_eval/models_configs/merlinite-7B-AOT/configs.yaml
@@ -5,7 +5,7 @@ merlinite-7B-AOT:
     model_name: "merlinite-7B-AOT"
     model_kwargs:
       torch_dtype: 'bfloat16'
-    do_sample: True
+    use_beam_search: True
     temperature: 0.7
     max_tokens: 2048
     top_p: 0.9
diff --git a/src/alpaca_eval/models_configs/openbuddy-llama2-13b-v11.1/configs.yaml b/src/alpaca_eval/models_configs/openbuddy-llama2-13b-v11.1/configs.yaml
index 515beeac..3f0d8355 100644
--- a/src/alpaca_eval/models_configs/openbuddy-llama2-13b-v11.1/configs.yaml
+++ b/src/alpaca_eval/models_configs/openbuddy-llama2-13b-v11.1/configs.yaml
@@ -9,7 +9,7 @@ openbuddy-llama2-13b-v11.1:
     max_new_tokens: 2000 # on vllm, prompt_tokens+max_new_tokens must be <= model_max_length otherwise it will raise an error and nothing will be generated
     temperature: 0.7
     top_p: 1.0
-    do_sample: False
+    use_beam_search: False
     batch_size: 16
   pretty_name: "OpenBudddy-LLaMA2-13B-v11.1"
   link: "https://huggingface.co/OpenBuddy/openbuddy-llama2-13b-v11.1-bf16"
diff --git a/src/alpaca_eval/models_configs/pairrm-tulu-2-13b/configs.yaml b/src/alpaca_eval/models_configs/pairrm-tulu-2-13b/configs.yaml
index 7bff9b9e..f39961be 100644
--- a/src/alpaca_eval/models_configs/pairrm-tulu-2-13b/configs.yaml
+++ b/src/alpaca_eval/models_configs/pairrm-tulu-2-13b/configs.yaml
@@ -9,7 +9,7 @@ pairrm-tulu-2-13b:
     max_new_tokens: 7500
     temperature: 0.0
     top_p: 1.0
-    do_sample: False
+    use_beam_search: False
     best_of: 16 # number of completions to generate, using PairRM to select the best one
     batch_size: 800
   pretty_name: "PairRM 0.4B+Tulu 2+DPO 13B (best-of-16)"
diff --git a/src/alpaca_eval/models_configs/pairrm-tulu-2-70b/configs.yaml b/src/alpaca_eval/models_configs/pairrm-tulu-2-70b/configs.yaml
index b63e9618..a3eb79e8 100644
--- a/src/alpaca_eval/models_configs/pairrm-tulu-2-70b/configs.yaml
+++ b/src/alpaca_eval/models_configs/pairrm-tulu-2-70b/configs.yaml
@@ -9,7 +9,7 @@ pairrm-tulu-2-70b:
     max_new_tokens: 7500
     temperature: 0.0
     top_p: 1.0
-    do_sample: False
+    use_beam_search: False
     best_of: 16 # number of completions to generate, using PairRM to select the best one
     batch_size: 800
   pretty_name: "PairRM 0.4B+Tulu 2+DPO 70B (best-of-16)"
diff --git a/src/alpaca_eval/models_configs/phi-2/configs.yaml b/src/alpaca_eval/models_configs/phi-2/configs.yaml
index 48bbe029..43679cb9 100644
--- a/src/alpaca_eval/models_configs/phi-2/configs.yaml
+++ b/src/alpaca_eval/models_configs/phi-2/configs.yaml
@@ -11,6 +11,6 @@ phi-2:
     max_new_tokens: 2048
     temperature: 0.7
     top_p: 1.0
-    do_sample: True
+    use_beam_search: True
   pretty_name: "Phi 2"
   link: "https://huggingface.co/microsoft/phi-2"
diff --git a/src/alpaca_eval/models_configs/tulu-2-dpo-13b-ExPO/configs.yaml b/src/alpaca_eval/models_configs/tulu-2-dpo-13b-ExPO/configs.yaml
index 68f32c8f..959d456e 100644
--- a/src/alpaca_eval/models_configs/tulu-2-dpo-13b-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/tulu-2-dpo-13b-ExPO/configs.yaml
@@ -7,7 +7,7 @@ tulu-2-dpo-13b-ExPO:
       dtype: 'bfloat16'
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/tulu-2-dpo-13b/configs.yaml b/src/alpaca_eval/models_configs/tulu-2-dpo-13b/configs.yaml
index f704b959..f9f6e2a4 100644
--- a/src/alpaca_eval/models_configs/tulu-2-dpo-13b/configs.yaml
+++ b/src/alpaca_eval/models_configs/tulu-2-dpo-13b/configs.yaml
@@ -9,7 +9,7 @@ tulu-2-dpo-13b:
     max_new_tokens: 7500
     temperature: 0.0
     top_p: 1.0
-    do_sample: False
+    use_beam_search: False
     batch_size: 800
   pretty_name: "Tulu 2+DPO 13B"
   link: "https://huggingface.co/allenai/tulu-2-dpo-13b"
\ No newline at end of file
diff --git a/src/alpaca_eval/models_configs/tulu-2-dpo-70b-ExPO/configs.yaml b/src/alpaca_eval/models_configs/tulu-2-dpo-70b-ExPO/configs.yaml
index 0057890c..c9bd1411 100644
--- a/src/alpaca_eval/models_configs/tulu-2-dpo-70b-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/tulu-2-dpo-70b-ExPO/configs.yaml
@@ -8,7 +8,7 @@ tulu-2-dpo-70b-ExPO:
       tokenizer_mode: 'auto'
       tensor_parallel_size: 2 # you need at least 2 A100 80GB GPUs to run this model
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/tulu-2-dpo-70b/configs.yaml b/src/alpaca_eval/models_configs/tulu-2-dpo-70b/configs.yaml
index 98de3280..431b2783 100644
--- a/src/alpaca_eval/models_configs/tulu-2-dpo-70b/configs.yaml
+++ b/src/alpaca_eval/models_configs/tulu-2-dpo-70b/configs.yaml
@@ -9,7 +9,7 @@ tulu-2-dpo-70b:
     max_new_tokens: 7500
     temperature: 0.0
     top_p: 1.0
-    do_sample: False
+    use_beam_search: False
     batch_size: 800
   pretty_name: "Tulu 2+DPO 70B"
   link: "https://huggingface.co/allenai/tulu-2-dpo-70b"
\ No newline at end of file
diff --git a/src/alpaca_eval/models_configs/tulu-2-dpo-7b-ExPO/configs.yaml b/src/alpaca_eval/models_configs/tulu-2-dpo-7b-ExPO/configs.yaml
index 1a26cae4..b703ef94 100644
--- a/src/alpaca_eval/models_configs/tulu-2-dpo-7b-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/tulu-2-dpo-7b-ExPO/configs.yaml
@@ -7,7 +7,7 @@ tulu-2-dpo-7b-ExPO:
       dtype: 'bfloat16'
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/tulu-2-dpo-7b/configs.yaml b/src/alpaca_eval/models_configs/tulu-2-dpo-7b/configs.yaml
index c4a55838..0cf00b9c 100644
--- a/src/alpaca_eval/models_configs/tulu-2-dpo-7b/configs.yaml
+++ b/src/alpaca_eval/models_configs/tulu-2-dpo-7b/configs.yaml
@@ -9,7 +9,7 @@ tulu-2-dpo-7b:
     max_new_tokens: 7500
     temperature: 0.0
     top_p: 1.0
-    do_sample: False
+    use_beam_search: False
     batch_size: 800
   pretty_name: "Tulu 2+DPO 7B"
   link: "https://huggingface.co/allenai/tulu-2-dpo-7b"
\ No newline at end of file
diff --git a/src/alpaca_eval/models_configs/zephyr-7b-alpha-ExPO/configs.yaml b/src/alpaca_eval/models_configs/zephyr-7b-alpha-ExPO/configs.yaml
index fd8763bb..9da03198 100644
--- a/src/alpaca_eval/models_configs/zephyr-7b-alpha-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/zephyr-7b-alpha-ExPO/configs.yaml
@@ -7,7 +7,7 @@ zephyr-7b-alpha-ExPO:
       dtype: 'bfloat16'
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50
diff --git a/src/alpaca_eval/models_configs/zephyr-7b-beta-ExPO/configs.yaml b/src/alpaca_eval/models_configs/zephyr-7b-beta-ExPO/configs.yaml
index 259df29f..33ab6da1 100644
--- a/src/alpaca_eval/models_configs/zephyr-7b-beta-ExPO/configs.yaml
+++ b/src/alpaca_eval/models_configs/zephyr-7b-beta-ExPO/configs.yaml
@@ -7,7 +7,7 @@ zephyr-7b-beta-ExPO:
       dtype: 'bfloat16'
       tokenizer_mode: 'auto'
     max_new_tokens: 2048
-    do_sample: True
+    use_beam_search: True
     seed: 42
     temperature: 0.7
     top_k: 50