Skip to content

fix(benchmark): disable thinking mode & improve JSON parsing#162

Merged
solderzzc merged 1 commit intodevelopfrom
feature/benchmark-thinking-mode-fix
Mar 17, 2026
Merged

fix(benchmark): disable thinking mode & improve JSON parsing#162
solderzzc merged 1 commit intodevelopfrom
feature/benchmark-thinking-mode-fix

Conversation

@solderzzc
Copy link
Copy Markdown
Member

  • Remove stream_options for local llama-server (causes crashes)
  • Drop max_tokens — streaming 2000-token cap is safety net
  • Enhance parseJSON for multi-word tags
  • Add JSON extraction fallback from reasoning_content
  • Simplify prompt template to avoid template echoing
  • Fix process.exit(1) in skill mode for clean status

- Remove stream_options for local llama-server (causes crashes)
- Drop max_tokens — streaming 2000-token cap is safety net
- Enhance parseJSON for multi-word <placeholder> tags
- Add JSON extraction fallback from reasoning_content
- Simplify prompt template to avoid template echoing
- Fix process.exit(1) in skill mode for clean status
@solderzzc solderzzc merged commit e9d7d4a into develop Mar 17, 2026
1 check passed
@solderzzc solderzzc deleted the feature/benchmark-thinking-mode-fix branch March 17, 2026 20:02
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant