Research map / Large language models / Language, safety and interpretability
Hallucination as a Computational Boundary: A Hierarchy of Inevitability and the Oracle Escape
Xi Wang, Quan Shi, Zenghui Ding, Jianqing Gao, Xianjun Yang
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| add continual-learning update to RAG (RAG-CL hybrid) | Pure RAG · Novel facts QA, Mistral-7B | amortized cost | Improved |
| add continual-learning update to RAG (RAG-CL hybrid) | Pure RAG · Novel facts QA, 15% noise | robustness: 76.5% -> 92.3% | Improved |
| add continual-learning update to RAG (RAG-CL hybrid) | Pure RAG · Novel facts QA | accuracy: ∼98.6% -> ∼96.5% | Hurt |
| add retrieval to LoRA continual learning (RAG-CL hybrid) | Pure CLM · Novel facts QA | accuracy: ∼81.0% -> ∼96.5% | Improved |
| add retrieval-triggered internalization to LoRA fine-tuning | Pure CLM · TriviaQA | forgetting rate: 12.4% -> 1.1% | Improved |
| internalize a fact via CL update | RAG-CL hybrid · Frequently queried fact | attention to RAG context | Improved |
Closest papers by meaning
- InEx: Hallucination Mitigation via Introspection and Cross-Modal Multi-Agent Collaboration (AAAI 2026 · similarity 0.83)
- Hallucination as Commitment Failure: Larger LLMs Misfire Despite Knowing the Answer (NeurIPS 2026 · similarity 0.83)
- Token-Guard: Towards Token-Level Hallucination Control via Self-Checking Decoding (ICLR 2026 · similarity 0.83)
- Can We Trust Embodied Agents? Exploring Backdoor Attacks against Embodied LLM-Based Decision-Making Systems (ICLR 2025 · similarity 0.83)
- Generating in the Limit with Infinitely Many Hallucinations (NeurIPS 2026 · similarity 0.83)
- BALTO: Balanced Token-Level Policy Optimization for Hallucination Mitigation (NeurIPS 2026 · similarity 0.82)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
