Back to corpus
proposalexperiment writeup candidatescore 40

RAG++ & Graph Kernel — Detailed Improvement Analysis

| Track | Target | Status | ETA | |-------|--------|--------|-----| | **Track 1: MiniMax Density Scoring** | kimi_memory.db (9.1K user turns) | 🟢 62% complete, 0 errors | ~1.5h | | **Track 2: RAG++ & Graph Kernel** | Search quality, entity enrichment | 🔴 Audited, issues documented | Needs work |

Open full HTMLDownload HTMLCorpus JSONPaper schema

Full HTML reader

Read the full artifact

Open in new tab

Extracted abstract or opening context

| Track | Target | Status | ETA | |-------|--------|--------|-----| | **Track 1: MiniMax Density Scoring** | kimi_memory.db (9.1K user turns) | 🟢 62% complete, 0 errors | ~1.5h | | **Track 2: RAG++ & Graph Kernel** | Search quality, entity enrichment | 🔴 Audited, issues documented | Needs work | **Running process:** PID 56430, `density_scorer.py --role user --min-length 50 --parallel 4` | Tier | Count | % | Description | |------|-------|---|-------------| | CORE | 146 | 2.6% | Deep values, unique decisions, strong preferences | | ENRICHED | 559 | 9.8% | Substantive context, project decisions | | ACTIVE | 2,983 | 52.3% | Normal productive conversation | | PRUNED | 1,317 | 23.1% | Low-value, boilerplate, noise | | ERROR (parse) | 695 | 12.2% | MiniMax output parsing failures | **Projected final (at current ratios):** - CORE: ~235 turns - ENRICHED: ~900 turns - **High-value (CORE+ENRICHED): ~1,135 turns** for v9 expansion **Output:** `Desktop/cognitive-twin/output/density_scores_20260216_183513.jsonl` (5,792 lines, 1.2MB)

Promotion decision

What has to happen next

Attach run IDs, datasets, metrics, and reproduction commands.

Why this is not always a full paper yet

Corpus pages are public-safe readers for discovered workspace artifacts. They are not automatically final papers. A corpus item becomes a polished paper only after the editable source, evidence checkpoints, references, figures, render path, and release status are attached through the paper schema.