-
Added prompt template config for system prompt and start sequence. 1 of 3 checklist items completed
- Merged
-
-
- 11
- Approved
updated -
Code creation pipeline 0 of 3 checklist items completed
- Merged
-
-
- 48
- Approved
updated -
Capture and write final prompt to output BQ table 1 of 3 checklist items completed
- Merged
-
-
- 3
- Approved
updated -
Add script to fetch latest metrics from daily run 2 of 3 checklist items completed
- Merged
-
-
- 18
- Approved
updated -
Add support to evaluate with Claude 3 Haiku 1 of 3 checklist items completed
- Merged
-
-
- 1
- Approved
updated -
Add Gemini Pro 1.5 model support 1 of 3 checklist items completed
- Merged
-
-
- 7
- Approved
updated -
Add claude-3 1 of 3 checklist items completed
- Merged
-
-
- 8
- Approved
updated -
Added ability to compare with ground truth 0 of 3 checklist items completed
- Merged
-
-
- 1
- Approved
updated -
Add pipeline to create code explanation dataset 1 of 3 checklist items completed
- Merged
-
-
- 30
- 1
- Approved
updated -
Add model handler from Gemini PRO 1 of 3 checklist items completed
- Merged
-
-
- 2
- Approved
updated -
Support local input to avoid calling GitLab API 2 of 3 checklist items completed
- Merged
-
-
- 27
- Approved
updated -
Add option to change the BigQuery write dispositon 3 of 3 checklist items completed
- Merged
-
-
- 2
- Approved
updated -
Add pipeline start timestamp 1 of 3 checklist items completed
- Merged
-
-
- 3
- Approved
updated -
Add LLM-judge binary metric 1 of 3 checklist items completed
-
Add pipeline to run Duo Chat evaluation locally without using GitLab API 2 of 3 checklist items completed
- Merged
-
- 3
updated -
Specify each metric separately in the config 1 of 3 checklist items completed
- Merged
-
-
- 14
- 1
- Approved
updated -
Support local output of Chat evaluation results 3 of 3 checklist items completed
- Merged
-
-
- 22
- Approved
updated -
Added dry run and test run for easy debugging 1 of 3 checklist items completed
- Merged
-
-
- 10
- Approved
updated -
Added text-bison-32k model 0 of 3 checklist items completed
- Merged
-
-
- Approved
updated