Query prompt: one space after the instruction, as used in the paper

README.md

@@ -43,9 +43,9 @@ before that) or `sentence-transformers>=3.0`.
 ## Usage
 
 The two sides are encoded asymmetrically: a task description carries an
-instruction prefix (no space between the prefix and the task), and a skill is its
-bare body. Encode them this way or retrieval quality drops. Embeddings come back
-L2-normalized, so cosine similarity is a plain dot product.
+instruction prefix followed by one space, and a skill is its bare body. Encode
+them this way or retrieval quality drops. Embeddings come back L2-normalized, so
+cosine similarity is a plain dot product.
 
 ```python
 import torch
@@ -58,7 +58,7 @@ model = AutoModel.from_pretrained(MODEL, dtype=torch.bfloat16).cuda().eval()
 
 QUERY_INSTRUCTION = (
     "Instruct: Given a coding task description, retrieve the most relevant "
-    "skill document that would help an agent complete the task\nQuery:"
+    "skill document that would help an agent complete the task\nQuery: "
 )
 
 

config_sentence_transformers.json

@@ -1,6 +1,6 @@
 {
   "prompts": {
-    "query": "Instruct: Given a coding task description, retrieve the most relevant skill document that would help an agent complete the task\nQuery:",
+    "query": "Instruct: Given a coding task description, retrieve the most relevant skill document that would help an agent complete the task\nQuery: ",
     "document": ""
   },
   "default_prompt_name": null,