Timothy-Vinzent commited on
Commit
c1cd0b6
·
verified ·
1 Parent(s): 3f8b483

Update app.py

Browse files
Files changed (1) hide show
  1. app.py +27 -3
app.py CHANGED
@@ -22,7 +22,7 @@ def get_evaluation_questions():
22
  Both lists must be of equal length.
23
  """
24
  questions_str = os.environ.get("TEST_QUESTION_1")
25
- expected_str = os.environ.get("TEST_EXPECTED")
26
  if not questions_str or not expected_str:
27
  return []
28
  try:
@@ -126,10 +126,34 @@ def build_interface():
126
  Constructs the Gradio interface with a submission button and single-submission mechanism.
127
  """
128
  with gr.Blocks() as demo:
129
- gr.Markdown("# GPT-4o Mini Prompt Submission")
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
130
  gr.Markdown(
131
  "Please enter your details and submit your system prompt below. "
132
- "You can only submit once."
133
  )
134
 
135
  email_input = gr.Textbox(label="Email", placeholder="your.email@example.com")
 
22
  Both lists must be of equal length.
23
  """
24
  questions_str = os.environ.get("TEST_QUESTION_1")
25
+ expected_str = os.environ.get("TEST_EXPECTED_1")
26
  if not questions_str or not expected_str:
27
  return []
28
  try:
 
126
  Constructs the Gradio interface with a submission button and single-submission mechanism.
127
  """
128
  with gr.Blocks() as demo:
129
+ gr.Markdown("# GPT-4o Mini System Prompt Submission")
130
+ gr.Markdown("""Classification Task: Document and Clause Level Identification
131
+ Challenge Description
132
+ Participants must create a system prompt for a language model that classifies user queries about legal documents into two specific categories:"
133
+
134
+ 1. Document Level: Determines whether the query refers to a single document or multiple documents.
135
+
136
+ 2. Clause Level: Identifies whether the query is focused on:
137
+
138
+ - A single clause,
139
+
140
+ - Multiple clauses, or
141
+
142
+ - General information not constrained to any specific clause.
143
+
144
+ The model must return a valid JSON object with the following structure:
145
+
146
+ json```
147
+ {
148
+ "document_level": "single/multiple",
149
+ "clause_level": "single/multiple/general"
150
+ }```
151
+ The goal is to ensure that the model's output is concise, structured, and accurate. This task is designed to evaluate the robustness of the system prompt in handling classification tasks with short, precise outputs.
152
+
153
+ """)
154
  gr.Markdown(
155
  "Please enter your details and submit your system prompt below. "
156
+ "You can only submit once, I suggest trying to test and build out the system prompt using the same LM being used here elsewhere before submitting."
157
  )
158
 
159
  email_input = gr.Textbox(label="Email", placeholder="your.email@example.com")