diff --git a/ci/run.sh b/ci/run.sh
index 7d241ecc0..7edae4f6e 100755
--- a/ci/run.sh
+++ b/ci/run.sh
@@ -751,7 +751,8 @@ function gg_run_rerank_tiny {
 
     model_f16="${path_models}/ggml-model-f16.gguf"
 
-    (time ./bin/llama-embedding --model ${model_f16}  -p "what is panda?</s><s>hi\nwhat is panda?</s><s>it's a bear\nwhat is panda?</s><s>The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China." --pooling rank --embd-normalize -1 --verbose-prompt) 2>&1 | tee -a $OUT/${ci}-rk-f16.log
+    # for this model, the SEP token is "</s>"
+    (time ./bin/llama-embedding --model ${model_f16}  -p "what is panda?</s></s>hi\nwhat is panda?</s></s>it's a bear\nwhat is panda?</s></s>The giant panda (Ailuropoda melanoleuca), sometimes called a panda bear or simply panda, is a bear species endemic to China." --pooling rank --embd-normalize -1 --verbose-prompt) 2>&1 | tee -a $OUT/${ci}-rk-f16.log
 
     # sample output
     # rerank score 0:    0.029
diff --git a/examples/server/server.cpp b/examples/server/server.cpp
index f343cc252..13e54e501 100644
--- a/examples/server/server.cpp
+++ b/examples/server/server.cpp
@@ -2027,7 +2027,7 @@ struct server_context {
                                 continue;
                             }
 
-                            // prompt: <s>query</s><s>doc</s>
+                            // prompt: [BOS]query[EOS][SEP]doc[EOS]
                             prompt_tokens.clear();
                             prompt_tokens.push_back(llama_token_bos(model));
                             {
@@ -2035,7 +2035,7 @@ struct server_context {
                                 prompt_tokens.insert(prompt_tokens.end(), part.begin(), part.end());
                             }
                             prompt_tokens.push_back(llama_token_eos(model));
-                            prompt_tokens.push_back(llama_token_bos(model));
+                            prompt_tokens.push_back(llama_token_sep(model));
                             {
                                 const auto part = tokenize(slot.prompt[1], false);
                                 prompt_tokens.insert(prompt_tokens.end(), part.begin(), part.end());