Research phase · F709

F709 — Remaining If Audit

F709 covers “Remaining If Audit”. The canonical record classifies it as failed/revert, based on 6 source artifacts and 2 original source figures. The highlighted measurements are Carryover Cases Passed = 6, Pass = Yes, and Decision = GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN.

F709 Failed/Revert GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN · KEEP_CANDIDATE

StatusFailed/Revert

VerdictGENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN · KEEP_CANDIDATE

Source figures2

Artifacts6

Multiple scoped outcomes. The source package records more than one outcome code. They remain visible, while the page status follows the most restrictive terminal signal and does not convert a partial or failed gate into a full pass.

Purpose and research question

The objective is to document the recorded outcome for “Remaining If Audit”. The research question is whether the phase-specific checks support the canonical failed/revert classification. The recorded measurements include Carryover Cases Passed = 6, Pass = Yes, and Decision = GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN.

Method

F709 is documented by 3 structured JSON records and 3 Markdown reports and 2 original source figures. The reported table retains safe scalar measurements found in that phase package.

Documentation basis

6 source artifacts and 2 phase-specific figures are recorded for this phase.

Evidence highlights

Carryover Cases Passed: 6
Pass: Yes
Decision: GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN
Metrics Memorized Span Ratio: 1
Metrics Novel Sequence Ratio: 0
Metrics Real Composition Rate: 0
Metrics Latency P50 Ms: 4322.147150014644
Metrics Latency P95 Ms: 5359.94870000286

Recorded evidence

Carryover Cases Passed	6
Pass	Yes
Decision	GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN
Metrics Memorized Span Ratio	1
Metrics Novel Sequence Ratio	0
Metrics Real Composition Rate	0
Metrics Latency P50 Ms	4322.147150014644
Metrics Latency P95 Ms	5359.94870000286

Public evidence and source figures

F709 benchmark evidence — Conversation Carryover Cleanup 709 — **Conversation Carryover Cleanup 709**
Original source figure · 1120 × 640 px

F709 benchmark evidence — Real Sentence Generator Anatomy 709 — **Real Sentence Generator Anatomy 709**
Original source figure · 2080 × 800 px

These are original source figures from the existing public evidence set, not editorial HTML visualizations.

Result and interpretation

F709 Remaining If Audit GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN · KEEP_CANDIDATE

F709 remains failed under the recorded decision GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN · KEEP_CANDIDATE. The recorded measurements include Carryover Cases Passed = 6, Pass = Yes, Decision = GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN, and Metrics Memorized Span Ratio = 1. These values define the negative phase result.

Key results

Carryover Cases Passed: 6
Pass: Yes
Decision: GENERATOR_EXISTS_BUT_CORE_COMPOSITION_IS_BROKEN
Metrics Memorized Span Ratio: 1
Metrics Novel Sequence Ratio: 0
Metrics Real Composition Rate: 0
Metrics Latency P50 Ms: 4322.147150014644
Metrics Latency P95 Ms: 5359.94870000286

Evidence sources

Values on this page are reproduced from the archived phase record and the source artifacts below. No new benchmark was run to create this page.

CONVERSATION CARRYOVER CLEANUP 709 (JSON)
CONVERSATION CARRYOVER CLEANUP 709 (MD)
REAL SENTENCE GENERATOR ANATOMY 709 (JSON)
REAL SENTENCE GENERATOR ANATOMY 709 (MD)
REMAINING IF AUDIT 709 (JSON)
REMAINING IF AUDIT 709 (MD)

Research topics

Word / World Document · Memory · Training · Benchmarks