WFGY/ProblemMap/GlobalFixMap/LLM_Providers/anthropic_claude.md
2025-09-05 10:58:24 +08:00

164 lines
10 KiB
Markdown
Raw Blame History

This file contains invisible Unicode characters

This file contains invisible Unicode characters that are indistinguishable to humans but may be processed differently by a computer. If you think that this is intentional, you can safely ignore this warning. Use the Escape button to reveal them.

# Anthropic Claude: Guardrails and Fix Patterns
<details>
<summary><strong>🧭 Quick Return to Map</strong></summary>
<br>
> You are in a sub-page of **LLM_Providers**.
> To reorient, go back here:
>
> - [**LLM_Providers** — model vendors and deployment options](./README.md)
> - [**WFGY Global Fix Map** — main Emergency Room, 300+ structured fixes](../README.md)
> - [**WFGY Problem Map 1.0** — 16 reproducible failure modes](../../README.md)
>
> 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.
</details>
**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](https://github.com/onestardao/WFGY/blob/main/ProblemMap/hallucination.md) and [Retrieval Playbook](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-playbook.md)
* Chunks look right but reasoning drifts or flips late
→ See [Retrieval Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-collapse.md) and [Logic Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/logic-collapse.md)
* High vector similarity yet wrong meaning
→ See [Embedding ≠ Semantic](https://github.com/onestardao/WFGY/blob/main/ProblemMap/embedding-vs-semantic.md)
* Very long context gets smeared, capitalization or topic wobble
→ See [Context Drift](https://github.com/onestardao/WFGY/blob/main/ProblemMap/context-drift.md) and [Entropy Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/entropy-collapse.md)
* Source mixing, who said what merges
→ See [Symbolic Constraint Unlock](https://github.com/onestardao/WFGY/blob/main/ProblemMap/patterns/pattern_symbolic_constraint_unlock.md) and [Retrieval Traceability](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-traceability.md)
* Fix seems to work then reverts next turn
→ See [Hallucination Re-entry](https://github.com/onestardao/WFGY/blob/main/ProblemMap/patterns/pattern_hallucination_reentry.md)
---
## Claude-specific gotchas and the repair
1. **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](https://github.com/onestardao/WFGY/blob/main/ProblemMap/logic-collapse.md) and [Multi-Agent Problems](https://github.com/onestardao/WFGY/blob/main/ProblemMap/Multi-Agent_Problems.md)
2. **System prompt anchoring beats your retrieval**
If the answer ignores citations, lock the schema: system → task → constraints → citations → answer.
Read: [Retrieval Traceability](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-traceability.md) and [Data Contracts](https://github.com/onestardao/WFGY/blob/main/ProblemMap/data-contracts.md)
3. **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](https://github.com/onestardao/WFGY/blob/main/ProblemMap/bluffing.md) and [Logic Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/logic-collapse.md)
4. **Very long window stability**
Gate OCR and chunk joins by ΔS, then stabilize attention with BBAM.
Read: [Entropy Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/entropy-collapse.md), [Hallucination](https://github.com/onestardao/WFGY/blob/main/ProblemMap/hallucination.md), and [Chunking Checklist](https://github.com/onestardao/WFGY/blob/main/ProblemMap/chunking-checklist.md)
5. **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](https://github.com/onestardao/WFGY/blob/main/ProblemMap/data-contracts.md) and [Retrieval Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-collapse.md)
---
## Minimal triage for Claude
1. **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](https://github.com/onestardao/WFGY/blob/main/ProblemMap/embedding-vs-semantic.md)
2. **Localize**
Tag λ across retrieval, assembly, reasoning.
* If λ flips only at reasoning, lock schema and apply BBAM
Read: [Retrieval Traceability](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-traceability.md) and [Logic Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/logic-collapse.md)
3. **Repair**
* Perception faults: semantic chunking, rebuild index with explicit metric, rerank if needed
Read: [Retrieval Playbook](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-playbook.md) and [Rerankers](https://github.com/onestardao/WFGY/blob/main/ProblemMap/rerankers.md)
* 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](https://github.com/onestardao/WFGY/blob/main/ProblemMap/rag-architecture-and-recovery.md)
---
## 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](https://github.com/onestardao/WFGY/blob/main/I_am_not_lizardman/WFGY_All_Principles_Return_to_One_v1.0_PSBigBig_Public.pdf) | 1⃣ Download · 2⃣ Upload to your LLM · 3⃣ Ask “Answer using WFGY + \<your question>” |
| **TXT OS (plain-text OS)** | [TXTOS.txt](https://github.com/onestardao/WFGY/blob/main/OS/TXTOS.txt) | 1⃣ Download · 2⃣ Paste into any LLM chat · 3⃣ Type “hello world” — OS boots instantly |
---
### 🧭 Explore More
| Module | Description | Link |
| ------------------------ | ---------------------------------------------------------------------------- | -------------------------------------------------------------------------------------------------- |
| WFGY Core | WFGY 2.0 engine is live: full symbolic reasoning architecture and math stack | [View →](https://github.com/onestardao/WFGY/tree/main/core/README.md) |
| Problem Map 1.0 | Initial 16-mode diagnostic and symbolic fix framework | [View →](https://github.com/onestardao/WFGY/tree/main/ProblemMap/README.md) |
| Problem Map 2.0 | RAG-focused failure tree, modular fixes, and pipelines | [View →](https://github.com/onestardao/WFGY/blob/main/ProblemMap/rag-architecture-and-recovery.md) |
| Semantic Clinic Index | Expanded failure catalog: prompt injection, memory bugs, logic drift | [View →](https://github.com/onestardao/WFGY/blob/main/ProblemMap/SemanticClinicIndex.md) |
| Semantic Blueprint | Layer-based symbolic reasoning & semantic modulations | [View →](https://github.com/onestardao/WFGY/tree/main/SemanticBlueprint/README.md) |
| Benchmark vs GPT-5 | Stress test GPT-5 with full WFGY reasoning suite | [View →](https://github.com/onestardao/WFGY/tree/main/benchmarks/benchmark-vs-gpt5/README.md) |
| 🧙‍♂️ Starter Village 🏡 | New here? Lost in symbols? Click here and let the wizard guide you through | [Start →](https://github.com/onestardao/WFGY/blob/main/StarterVillage/README.md) |
---
> 👑 **Early Stargazers: [See the Hall of Fame](https://github.com/onestardao/WFGY/tree/main/stargazers)** —
> Engineers, hackers, and open source builders who supported WFGY from day one.
> <img src="https://img.shields.io/github/stars/onestardao/WFGY?style=social" alt="GitHub stars"> ⭐ [WFGY Engine 2.0](https://github.com/onestardao/WFGY/blob/main/core/README.md) is already unlocked. ⭐ Star the repo to help others discover it and unlock more on the [Unlock Board](https://github.com/onestardao/WFGY/blob/main/STAR_UNLOCKS.md).
<div align="center">
[![WFGY Main](https://img.shields.io/badge/WFGY-Main-red?style=flat-square)](https://github.com/onestardao/WFGY)
 
[![TXT OS](https://img.shields.io/badge/TXT%20OS-Reasoning%20OS-orange?style=flat-square)](https://github.com/onestardao/WFGY/tree/main/OS)
 
[![Blah](https://img.shields.io/badge/Blah-Semantic%20Embed-yellow?style=flat-square)](https://github.com/onestardao/WFGY/tree/main/OS/BlahBlahBlah)
 
[![Blot](https://img.shields.io/badge/Blot-Persona%20Core-green?style=flat-square)](https://github.com/onestardao/WFGY/tree/main/OS/BlotBlotBlot)
 
[![Bloc](https://img.shields.io/badge/Bloc-Reasoning%20Compiler-blue?style=flat-square)](https://github.com/onestardao/WFGY/tree/main/OS/BlocBlocBloc)
 
[![Blur](https://img.shields.io/badge/Blur-Text2Image%20Engine-navy?style=flat-square)](https://github.com/onestardao/WFGY/tree/main/OS/BlurBlurBlur)
 
[![Blow](https://img.shields.io/badge/Blow-Game%20Logic-purple?style=flat-square)](https://github.com/onestardao/WFGY/tree/main/OS/BlowBlowBlow)
 
</div>