allenpark commited on
Commit
5f886cd
1 Parent(s): e69d713

fix: update docs link

Browse files
Files changed (1) hide show
  1. app.py +1 -1
app.py CHANGED
@@ -166,7 +166,7 @@ HEADER = """
166
  <br>
167
  <div style="color: #fff !important; width: 70%"><span style="color: inherit; font-weight: 600">GLIDER</span> is a powerful 3B evaluator LLM that can score any text input and associated context on arbitrary user defined criteria.</div>
168
  <br>
169
- <div style="color: #fff !important; width: 70%;"><span style="color: inherit; font-weight: 600">Getting Started</span>: First, provide a model output (text generated by your model) and user input (text used to prompt your model) and optionally a gold answer (label or gold answer to the prompt) and retrieved context (context used for text generated by your model). Next, provide a pass criteria (description of a passing evaluation). Finally, provide an optional but recommended rubric (scoring scales with explanations) and then click submit. The GLIDER Output panel will provide a score and reasoning which is a human readable explanation of the score. You can find our docs <a href="https://docs.patronus.ai/docs/glider-research">here</a>.</div>
170
  <br>
171
  <div style="color: #fff !important; width: 70%;"><span style="color: inherit; font-weight: 600">Note</span>: In your Pass Criteria, use the attribute tags from above (i.e. MODEL OUTPUT, USER INPUT, GOLD ANSWER, RETRIEVED CONTEXT).</div>
172
  <br>
 
166
  <br>
167
  <div style="color: #fff !important; width: 70%"><span style="color: inherit; font-weight: 600">GLIDER</span> is a powerful 3B evaluator LLM that can score any text input and associated context on arbitrary user defined criteria.</div>
168
  <br>
169
+ <div style="color: #fff !important; width: 70%;"><span style="color: inherit; font-weight: 600">Getting Started</span>: First, provide a model output (text generated by your model) and user input (text used to prompt your model) and optionally a gold answer (label or gold answer to the prompt) and retrieved context (context used for text generated by your model). Next, provide a pass criteria (description of a passing evaluation). Finally, provide an optional but recommended rubric (scoring scales with explanations) and then click submit. The GLIDER Output panel will provide a score and reasoning which is a human readable explanation of the score. You can find our docs <a href="https://docs.patronus.ai/docs/evals-with-glider">here</a>.</div>
170
  <br>
171
  <div style="color: #fff !important; width: 70%;"><span style="color: inherit; font-weight: 600">Note</span>: In your Pass Criteria, use the attribute tags from above (i.e. MODEL OUTPUT, USER INPUT, GOLD ANSWER, RETRIEVED CONTEXT).</div>
172
  <br>