nx2/test-small-llms
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

typos

Browse Source
This commit is contained in:
Lennart J. Kurzweg (Nx2)
2024-10-05 20:47:56 +02:00
parent ae2ea6d3c9
commit 314077a63d
3 changed files with 6 additions and 6 deletions
Download Patch File Download Diff File Expand all files Collapse all files

2
libs/runnables.py
View File

@@ -220,7 +220,7 @@ def agent_with_tools(model: Model, seed: int, test: Test, technique: Technique,
chunks.append(chunk["messages"][-1]) chunks.append(chunk["messages"][-1])
except RecursionError: except RecursionError:
return { return {
"answer": ">>Model did not come to a conclusion (Recusion Error)<<", "answer": ">>Model did not come to a conclusion (Recursion Error)<<",
"tool_calls": tool_calls "tool_calls": tool_calls
} }

4
libs/tools.py
View File

@@ -163,7 +163,7 @@ def write_note(note: str) -> str:
@tool @tool
def save_python_repl(command: str): def save_python_repl(command: str):
"""Simulates the normal python repl but with certain patterns blocked for savety reasons""" """Simulates the normal python repl but with certain patterns blocked for safety reasons"""
python_repl = PythonREPL() python_repl = PythonREPL()
blocked_patterns = [ blocked_patterns = [
"^ *os\\.", "^ *os\\.",
@@ -187,7 +187,7 @@ def save_python_repl(command: str):
return ret return ret
else: else:
return f"Command not executed, becaise the blocked pattern `{pattern}` was found in the command." return f"Command not executed, because the blocked pattern `{pattern}` was found in the command."

6
suite_settings/tests.py
View File

@@ -73,7 +73,7 @@ tests = {
validator=system_human_answer_match, validator=system_human_answer_match,
validation_input={ validation_input={
"criteria": dedent("""- containing the information that the Human should call Wolfgang "criteria": dedent("""- containing the information that the Human should call Wolfgang
- just one single conversational answer, without any AI fragments (A/B versions, "end of message" parts, unfitting discalimers or notes, what specific tool was used to get the answer, etc.)""") - just one single conversational answer, without any AI fragments (A/B versions, "end of message" parts, unfitting disclaimers or notes, what specific tool was used to get the answer, etc.)""")
}, },
), ),
260: Test( 260: Test(
@@ -91,7 +91,7 @@ tests = {
validator=system_human_answer_match, validator=system_human_answer_match,
validation_input={ validation_input={
"criteria": dedent("""- containing the information that the Human should call Wolfgang "criteria": dedent("""- containing the information that the Human should call Wolfgang
- just one single conversational answer, without any AI fragments (A/B versions, "end of message" parts, unfitting discalimers or notes, what specific tool was used to get the answer, etc.)""") - just one single conversational answer, without any AI fragments (A/B versions, "end of message" parts, unfitting disclaimers or notes, what specific tool was used to get the answer, etc.)""")
}, },
), ),
856: Test( 856: Test(
@@ -124,7 +124,7 @@ tests = {
validator=system_human_answer_match, validator=system_human_answer_match,
validation_input={ validation_input={
"criteria": dedent("""- containing the information that the Human should call Wolfgang "criteria": dedent("""- containing the information that the Human should call Wolfgang
- just one single conversational answer, without any AI fragments (A/B versions, "end of message" parts, unfitting discalimers or notes, what specific tool was used to get the answer, etc.)""") - just one single conversational answer, without any AI fragments (A/B versions, "end of message" parts, unfitting disclaimers or notes, what specific tool was used to get the answer, etc.)""")
}, },
), ),
# 600: Test(), # 600: Test(),