kami 1890ff5d5d Constrain the phrasing output with a GBNF grammar ...

The 0.8B answered about one chat turn in three with open reasoning as plain text, so no JSON ever closed and the fallback shipped "Thinking Process:" to the user. A grammar makes that output impossible.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CGeSZxh1DCtRxmFVSYVGvJ

2026-07-31 17:18:06 +04:00

..

eval

Report every address break, and stop -те verbs blinding the check

2026-07-31 16:52:16 +04:00

context_block_test.go

Give every prompt one shared context block

2026-07-31 15:55:30 +04:00

grammar_test.go

Constrain the phrasing output with a GBNF grammar

2026-07-31 17:18:06 +04:00

llmphraser.go

Constrain the phrasing output with a GBNF grammar

2026-07-31 17:18:06 +04:00

phraser_test.go

feat: LLM phraser, shared LLM client, and LLM replier

2026-07-10 15:48:55 +04:00

phraser.go

Add a deterministic scorer for nudge phrasing (Vikunja #323)

2026-07-31 02:30:52 +04:00