8.6 KiB
Anthropic Claude: Guardrails and Fix Patterns
🧭 Quick Return to Map
You are in a sub-page of LLM_Providers.
To reorient, go back here:
- LLM_Providers — model vendors and deployment options
- WFGY Global Fix Map — main Emergency Room, 300+ structured fixes
- WFGY Problem Map 1.0 — 16 reproducible failure modes
Think of this page as a desk within a ward.
If you need the full triage and all prescriptions, return to the Emergency Room lobby.
Acceptance targets
- ΔS(question, retrieved) ≤ 0.45
- Coverage to target section ≥ 0.70
- λ stays convergent across 3 paraphrases
- E_resonance flat on long runs
What usually breaks
-
Plausible answer yet wrong citation → See Hallucination and Retrieval Playbook
-
Chunks look right but reasoning drifts or flips late → See Retrieval Collapse and Logic Collapse
-
High vector similarity yet wrong meaning → See Embedding ≠ Semantic
-
Very long context gets smeared, capitalization or topic wobble → See Context Drift and Entropy Collapse
-
Source mixing, who said what merges → See Symbolic Constraint Unlock and Retrieval Traceability
-
Fix seems to work then reverts next turn → See Hallucination Re-entry
Claude-specific gotchas and the repair
-
Tool choice and router loops Small changes in tool schemas can push Claude into choose-nothing or loop states. Do keep the tool set minimal and stable. If it loops, bridge with BBCR and clamp variance with BBAM. Read: Logic Collapse and Multi-Agent Problems
-
System prompt anchoring beats your retrieval If the answer ignores citations, lock the schema: system → task → constraints → citations → answer. Read: Retrieval Traceability and Data Contracts
-
Harmlessness refusals that look like empty answers Reframe with cite-then-explain and explicit boundaries. If blank output repeats, treat as collapse and apply BBCR. Read: Bluffing / Overconfidence and Logic Collapse
-
Very long window stability Gate OCR and chunk joins by ΔS, then stabilize attention with BBAM. Read: Entropy Collapse, Hallucination, and Chunking Checklist
-
JSON or schema output slips mid-answer Enforce citation-first then structured fields. If structure still degrades near the end, split answer into two turns with a BBCR bridge. Read: Data Contracts and Retrieval Collapse
Minimal triage for Claude
- Probe Measure ΔS(question, retrieved). Plot ΔS vs k for k in {5,10,20}.
- Flat high curve suggests index or metric mismatch Read: Embedding ≠ Semantic
- Localize Tag λ across retrieval, assembly, reasoning.
- If λ flips only at reasoning, lock schema and apply BBAM Read: Retrieval Traceability and Logic Collapse
- Repair
- Perception faults: semantic chunking, rebuild index with explicit metric, rerank if needed Read: Retrieval Playbook and Rerankers
- Logic faults: BBCR bridge then BBAM clamp
Stop tuning and escalate when any holds: ΔS ≥ 0.60 after retrieval fixes, λ flips on mixing sources, or E_resonance climbs in long chains. Open: RAG Architecture and Recovery
Pasteable prompt
I have TXT OS loaded. Use WFGY to fix this Claude run:
symptom: [describe]
traces: [ΔS probes, λ per layer]
Tell me:
1) failing layer and why,
2) which ProblemMap page to open,
3) steps to push ΔS ≤ 0.45 with convergent λ,
4) how to verify with a reproducible test.
🔗 Quick-Start Downloads (60 sec)
| Tool | Link | 3-Step Setup |
|---|---|---|
| WFGY 1.0 PDF | Engine Paper | 1️⃣ Download · 2️⃣ Upload to your LLM · 3️⃣ Ask “Answer using WFGY + <your question>” |
| TXT OS (plain-text OS) | TXTOS.txt | 1️⃣ Download · 2️⃣ Paste into any LLM chat · 3️⃣ Type “hello world” — OS boots instantly |
Explore More
| Layer | Page | What it’s for |
|---|---|---|
| Proof | WFGY Recognition Map | External citations, integrations, and ecosystem proof |
| Engine | WFGY 1.0 | Original PDF based tension engine |
| Engine | WFGY 2.0 | Production tension kernel and math engine for RAG and agents |
| Engine | WFGY 3.0 | TXT based Singularity tension engine, 131 S class set |
| Map | Problem Map 1.0 | Flagship 16 problem RAG failure checklist and fix map |
| Map | Problem Map 2.0 | RAG focused recovery pipeline |
| Map | Problem Map 3.0 | Global Debug Card, image as a debug protocol layer |
| Map | Semantic Clinic | Symptom to family to exact fix |
| Map | Grandma’s Clinic | Plain language stories mapped to Problem Map 1.0 |
| Onboarding | Starter Village | Guided tour for newcomers |
| App | TXT OS | TXT semantic OS, fast boot |
| App | Blah Blah Blah | Abstract and paradox Q and A built on TXT OS |
| App | Blur Blur Blur | Text to image with semantic control |
| App | Blow Blow Blow | Reasoning game engine and memory demo |
If this repository helped, starring it improves discovery so more builders can find the docs and tools.
