Skip to content

AutoGPT is the vision of accessible AI for everyone, to use and to build on. Our mission is to provide the tools, so that you can focus on what matters.

fix(blocks): Fix feedback loops in AI Structured Response Generator (#10932)

Reinier van der Leerc2f11dbcfa3c
diff --git a/autogpt_platform/backend/backend/blocks/llm.py b/autogpt_platform/backend/backend/blocks/llm.py
index b20df4052e..856d736dcd 100644
--- a/autogpt_platform/backend/backend/blocks/llm.py
+++ b/autogpt_platform/backend/backend/blocks/llm.py
@@ -896,6 +896,7 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
prompt = [json.to_dict(p) for p in input_data.conversation_history]
def trim_prompt(s: str) -> str:
+ """Removes indentation up to and including `|` from a multi-line prompt."""
lines = s.strip().split("\n")
return "\n".join([line.strip().lstrip("|") for line in lines])
@@ -909,24 +910,25 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
if input_data.expected_format:
expected_format = [
- f'"{k}": "{v}"' for k, v in input_data.expected_format.items()
+ f"{json.dumps(k)}: {json.dumps(v)}"
+ for k, v in input_data.expected_format.items()
]
if input_data.list_result:
format_prompt = (
f'"results": [\n {{\n {", ".join(expected_format)}\n }}\n]'
)
else:
- format_prompt = "\n ".join(expected_format)
+ format_prompt = ",\n| ".join(expected_format)
sys_prompt = trim_prompt(
f"""
- |Reply strictly only in the following JSON format:
- |{{
- | {format_prompt}
- |}}
- |
- |Ensure the response is valid JSON. Do not include any additional text outside of the JSON.
- |If you cannot provide all the keys, provide an empty string for the values you cannot answer.
+ |Reply with pure JSON strictly following this JSON format:
+ |{{
+ | {format_prompt}
+ |}}
+ |
+ |Ensure the response is valid JSON. DO NOT include any additional text (e.g. markdown code block fences) outside of the JSON.
+ |If you cannot provide all the keys, provide an empty string for the values you cannot answer.
"""
)
prompt.append({"role": "system", "content": sys_prompt})
@@ -946,7 +948,7 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
return f"JSON decode error: {e}"
logger.debug(f"LLM request: {prompt}")
- retry_prompt = ""
+ error_feedback_message = ""
llm_model = input_data.model
for retry_count in range(input_data.retry):
@@ -970,8 +972,25 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
logger.debug(f"LLM attempt-{retry_count} response: {response_text}")
if input_data.expected_format:
-
- response_obj = json.loads(response_text)
+ try:
+ response_obj = json.loads(response_text)
+ except JSONDecodeError as json_error:
+ prompt.append({"role": "assistant", "content": response_text})
+
+ indented_json_error = str(json_error).replace("\n", "\n|")
+ error_feedback_message = trim_prompt(
+ f"""
+ |Your previous response could not be parsed as valid JSON:
+ |
+ |{indented_json_error}
+ |
+ |Please provide a valid JSON response that matches the expected format.
+ """
+ )
+ prompt.append(
+ {"role": "user", "content": error_feedback_message}
+ )
+ continue
if input_data.list_result and isinstance(response_obj, dict):
if "results" in response_obj:
@@ -979,7 +998,7 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
elif len(response_obj) == 1:
response_obj = list(response_obj.values())
- response_error = "\n".join(
+ validation_errors = "\n".join(
[
validation_error
for response_item in (
@@ -991,7 +1010,7 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
]
)
- if not response_error:
+ if not validation_errors:
self.merge_stats(
NodeExecutionStats(
llm_call_count=retry_count + 1,
@@ -1001,6 +1020,16 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
yield "response", response_obj
yield "prompt", self.prompt
return
+
+ prompt.append({"role": "assistant", "content": response_text})
+ error_feedback_message = trim_prompt(
+ f"""
+ |Your response did not match the expected format:
+ |
+ |{validation_errors}
+ """
+ )
+ prompt.append({"role": "user", "content": error_feedback_message})
else:
self.merge_stats(
NodeExecutionStats(
@@ -1011,21 +1040,6 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
yield "response", {"response": response_text}
yield "prompt", self.prompt
return
-
- retry_prompt = trim_prompt(
- f"""
- |This is your previous error response:
- |--
- |{response_text}
- |--
- |
- |And this is the error:
- |--
- |{response_error}
- |--
- """
- )
- prompt.append({"role": "user", "content": retry_prompt})
except Exception as e:
logger.exception(f"Error calling LLM: {e}")
if (
@@ -1038,9 +1052,12 @@ class AIStructuredResponseGeneratorBlock(AIBlockBase):
logger.debug(
f"Reducing max_tokens to {input_data.max_tokens} for next attempt"
)
- retry_prompt = f"Error calling LLM: {e}"
+ # Don't add retry prompt for token limit errors,
+ # just retry with lower maximum output tokens
+
+ error_feedback_message = f"Error calling LLM: {e}"
- raise RuntimeError(retry_prompt)
+ raise RuntimeError(error_feedback_message)
class AITextGeneratorBlock(AIBlockBase):