This table compares the performance of different edit formats (diff-fenced and CEDARScript) for various benchmark tasks in Gemini 1.5 PRO conversations.
| Task | Edit Format | Lines | Result |
|---|---|---|---|
| doc_DocCLI_get_role_man_text | diff-fenced | 4948 | FAILED (failures=1) |
| doc_DocCLI_get_role_man_text | CEDARScript | 45 | OK |
| analyzer_cli_DebugAnalyzer__make_source_table | diff-fenced | 4497 | FAILED (failures=1) |
| analyzer_cli_DebugAnalyzer__make_source_table | CEDARScript | 164 | OK |