WFGY/ProblemMap/GlobalFixMap/RAG/hallucination_rag.md
2025-09-05 11:40:30 +08:00

132 lines
8.3 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.

# Hallucination in RAG — Guardrails and Fix Pattern
<details>
<summary><strong>🧭 Quick Return to Map</strong></summary>
<br>
> You are in a sub-page of **RAG**.
> To reorient, go back here:
>
> - [**RAG** — retrieval-augmented generation and knowledge grounding](./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>
When the retriever returns plausible text, but the LLM answers with facts that are not in the corpus.
This page stabilizes retrieval → reasoning boundaries and applies WFGY symbolic fixes to stop hallucination loops.
---
## Open these first
- Visual recovery: [RAG Architecture & Recovery](https://github.com/onestardao/WFGY/blob/main/ProblemMap/rag-architecture-and-recovery.md)
- Retrieval schema: [Retrieval Traceability](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-traceability.md)
- Snippet contracts: [Data Contracts](https://github.com/onestardao/WFGY/blob/main/ProblemMap/data-contracts.md)
- Boundary fences: [Cite-then-Explain](https://github.com/onestardao/WFGY/blob/main/ProblemMap/cite-then-explain.md)
- Chain stability: [Logic Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/logic-collapse.md)
- Long-context drift: [Context Drift](https://github.com/onestardao/WFGY/blob/main/ProblemMap/context-drift.md)
---
## Core acceptance
- Every answer includes citation fields (`snippet_id`, `section_id`, `source_url`)
- ΔS(answer, cited snippet) ≤ 0.45
- Coverage ≥ 0.70 for cited target section
- λ convergent across 3 paraphrases and 2 seeds
---
## Typical symptoms → exact fix
| Symptom | Likely cause | Open this |
|---------|--------------|-----------|
| Answer has no citation or cites wrong section | missing contract schema | [Data Contracts](https://github.com/onestardao/WFGY/blob/main/ProblemMap/data-contracts.md), [Retrieval Traceability](https://github.com/onestardao/WFGY/blob/main/ProblemMap/retrieval-traceability.md) |
| LLM fabricates plausible filler | boundary collapse between retrieval and reasoning | [Cite-then-Explain](https://github.com/onestardao/WFGY/blob/main/ProblemMap/cite-then-explain.md), [Logic Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/logic-collapse.md) |
| Correct fact retrieved, but answer overwrites it | λ divergence in long chain | [Context Drift](https://github.com/onestardao/WFGY/blob/main/ProblemMap/context-drift.md), [Entropy Collapse](https://github.com/onestardao/WFGY/blob/main/ProblemMap/entropy-collapse.md) |
| Hallucination reappears after correction | unstable feedback loop | [Pattern: Hallucination Re-entry](https://github.com/onestardao/WFGY/blob/main/ProblemMap/patterns/pattern_hallucination_reentry.md) |
---
## Fix in 60 seconds
1. **Cite-first enforcement**
- Force the LLM to output citations before free text.
- Forbid answers that lack at least one valid snippet_id.
2. **ΔS probe**
- Compute ΔS(answer, cited snippet).
- If ≥ 0.60, trigger BBPF alternate reasoning path.
3. **Apply module**
- Boundary collapse → BBCR bridge + [Cite-then-Explain](https://github.com/onestardao/WFGY/blob/main/ProblemMap/cite-then-explain.md)
- Re-entry loops → [Pattern: Hallucination Re-entry](https://github.com/onestardao/WFGY/blob/main/ProblemMap/patterns/pattern_hallucination_reentry.md)
---
## Copy-paste prompt for debug
```txt
I uploaded TXT OS and the WFGY Problem Map.
My issue:
- retrieval finds the fact, but answer hallucinates new content
- traces: ΔS(answer,cited_snippet)=..., λ across 3 paraphrases
Tell me:
1. failing boundary (retrieval vs reasoning),
2. exact WFGY page to open,
3. minimal fix to enforce cite-then-explain and push ΔS ≤ 0.45,
4. test to verify stability over 3 paraphrases.
````
---
### 🔗 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.
> ⭐ [WFGY Engine 2.0](https://github.com/onestardao/WFGY/blob/main/core/README.md) is already unlocked. ⭐
<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>