arborist/aborist/qa
russell@unturf.com 38cfea1983
qa(verify): FORMAT_COLLAPSED soft-demote + open #000008 (broad-quantifier preflight)
Sister rule to Rule 9 (SUBJECT_TOKENS_ABSENT) landed in the same
session. Both demote STRICT → HYBRID but on orthogonal signals:
Rule 9 catches premise-parroting; FORMAT_COLLAPSED catches
protocol abandonment.

Surfaced by fox's "winners of all major sports?" 2026-05-02 case:
Hermes-3-8B melted under an under-specified broad-quantifier
question, dumped 50+ free-form prose claims with zero [E\d+]
pointer tags. Verifier honestly returned UNGROUNDED 0/2 (parser
caught two line fragments), but operators couldn't distinguish
"tried & failed to ground" from "abandoned the protocol." This
soft-demote separates the two failure shapes at audit-line glance.

verify_claim_lattice (pointer-mode only — JSON collapse already
shows as SCHEMA_INVALID):
- count meaningful_lines (>20 chars after strip) and [E\d+ regex
  matches in raw answer
- ≥5 meaningful lines AND 0 bracket tags → FORMAT_COLLAPSED
  violation, soft-demote STRICT → HYBRID
- format_collapsed: bool added to verdict dict

Plumbing:
- claim_lattice_format_collapse_check_enabled: True in DEFAULT_POLICY
  and DEFAULT_QUERY_POLICY
- _VERIFIER_POLICY_FIELDS in keys.py adds the field so it folds
  into verifier_policy_hash
- threaded through ask() and query() call sites

CLI:
- _SOFT_DEMOTE_VIOLATION_KINDS includes FORMAT_COLLAPSED so the
  audit-line ladder rendering treats it as a soft demote
- _render_warrant_tail appends "· format collapsed" tail

Bench fixture: new "under-specified 'all'" section in
qa_questions.txt with `winners of all major sports?` and rationale
about cross-model resilience signal.

Tests:
- test_format_collapsed_fires_on_bracketless_multi_line_prose
- test_format_collapsed_does_not_fire_when_pointer_tags_present
- CLI render coverage
Full suite: 781 passed (up from 776).

Open Ticket #000008 — Broad-quantifier preflight guard. Cleaner
upstream fix: detect quantifier-intensity at query layer and
apply a per-model claim ceiling BEFORE the 13-second LLM call.
FORMAT_COLLAPSED stays as the downstream catch; #000008 proposes
the upstream prevention. TICKETS.md index + Next ID 000008→000009.
2026-05-02 16:43:41 -04:00
..
__init__.py multi-source corpus query: pose a question, the tree pulls related cached docs 2026-04-27 11:43:57 -04:00
client.py qa(client): scrub lone UTF-16 surrogates before httpx encode 2026-05-02 11:58:25 -04:00
concepts.py qa(concepts): corpus-derived concept_relations table replaces frozensets 2026-05-01 21:14:53 -04:00
dag.py speed: pytest-xdist, bench smoke, concurrency default; UTF surrogate fix 2026-05-02 09:29:40 -04:00
evidence.py speed: pytest-xdist, bench smoke, concurrency default; UTF surrogate fix 2026-05-02 09:29:40 -04:00
frame.py docs: tickets/ subdir + delete naming-deferral, ref-frame, coverage-audit 2026-05-02 12:51:09 -04:00
inspect.py qa(inspect): wordlist union (US+UK) + configurable supplemental dictionaries 2026-05-02 14:27:24 -04:00
keys.py qa(verify): FORMAT_COLLAPSED soft-demote + open #000008 (broad-quantifier preflight) 2026-05-02 16:43:41 -04:00
parse_claims.py qa: G0 claim-lattice-pointer answer mode — CTI quote-by-pointer 2026-04-29 21:19:19 -04:00
prompts.py docs(bench): rename qa-modes-bench-2026-05-02.md → qa-modes-bench.md 2026-05-02 12:27:38 -04:00
query.py qa(verify): FORMAT_COLLAPSED soft-demote + open #000008 (broad-quantifier preflight) 2026-05-02 16:43:41 -04:00
repair.py qa: chain-segment failure localization + re-prompt repair tier 2026-04-29 19:27:33 -04:00
retrieval_plan.py docs: tickets/ subdir + delete naming-deferral, ref-frame, coverage-audit 2026-05-02 12:51:09 -04:00
runner.py qa(verify): FORMAT_COLLAPSED soft-demote + open #000008 (broad-quantifier preflight) 2026-05-02 16:43:41 -04:00
verify.py qa(verify): FORMAT_COLLAPSED soft-demote + open #000008 (broad-quantifier preflight) 2026-05-02 16:43:41 -04:00
warrant.py docs: tickets/ subdir + delete naming-deferral, ref-frame, coverage-audit 2026-05-02 12:51:09 -04:00