kami/Maven
1
0
Fork 0
Code Issues 1 Pull Requests 1 Actions Packages Projects Releases Wiki Activity

Stop shipping half-written JSON as a reply #42

Closed
claude wants to merge 2 commits from overnight/fix-truncation into overnight/ru-prompts
pull from: overnight/fix-truncation
merge into: kami:overnight/ru-prompts
Conversation 1 Commits 2 Files Changed 5
+170 -19

2 Commits

Author SHA1 Message Date
kami aa8f5b2ee2 Make the nonempty check look for actual words 
It scored 27/27 on a run where two replies were "{" and "{\n  \"". It only
tested that the string was not blank, so punctuation counted as content and
the worst replies of the run passed the first check.

Now a reply needs at least one letter, Cyrillic or Latin. Latin counts
because answers about ssd or vpn are legitimately part English.

Digits alone fail too. The same run answered "сколько варить яйцо
вкрутую?" with "15-16" — no unit, no words, and the wrong number as well.
That is not something she said.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CGeSZxh1DCtRxmFVSYVGvJ
 2026-07-31 17:57:39 +04:00
kami d7cdcb63bd Stop shipping half-written JSON as a reply 
Two bugs, one symptom. A run of the talk eval produced replies that were
literally "{" and "{\n  \"" — those strings went out as things Maven said.

First bug: the parser could not tell "the model answered in plain prose"
from "the model started a JSON object and got cut off". Both came back as
empty, and every caller then shipped the raw text. Now an unfinished object
returns an error and each caller uses its own fallback instead. Bare prose
with no JSON in it still passes through, because small models do sometimes
answer that way and the reply is fine.

Second bug, and the actual cause: the grammar capped the response field at
400 characters. I measured it against Qwen3.5-0.8B at three different token
caps — 256, 768 and 2048 — and the reply came back exactly 400 characters
every time, cut mid-word. So the token limit was never what stopped it.
The bound is 1000 now, about six Russian sentences, still low enough to cut
off a repetition loop.

Token caps go from 256 to 768 on the chat and query paths so 1000
characters of Russian actually fits. The nudge path keeps its own cap; a
nudge is meant to be one sentence.

Note: cmd/mavend/replier_llm.go has its own copy of this parser with the
same bug. Left alone here so this commit stays small — that duplicate is
Vikunja #396.

Co-Authored-By: Claude Opus 5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01CGeSZxh1DCtRxmFVSYVGvJ
 2026-07-31 17:56:55 +04:00