Maintaining Informative Coherence: Migrating Hallucinations in Large Language Models via Absorbing Markov Chains
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2024
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866929561412829184 |
|---|---|
| author | Wu, Jiemin Lai, Songning Xiao, Ruiqiang Xue, Tianlang Yang, Jiayu Yue, Yutao |
| author_facet | Wu, Jiemin Lai, Songning Xiao, Ruiqiang Xue, Tianlang Yang, Jiayu Yue, Yutao |
| contents | Large Language Models (LLMs) are powerful tools for text generation, translation, and summarization, but they often suffer from hallucinations-instances where they fail to maintain the fidelity and coherence of contextual information during decoding, sometimes overlooking critical details due to their sampling strategies and inherent biases from training data and fine-tuning discrepancies. These hallucinations can propagate through the web, affecting the trustworthiness of information disseminated online. To address this issue, we propose a novel decoding strategy that leverages absorbing Markov chains to quantify the significance of contextual information and measure the extent of information loss during generation. By considering all possible paths from the first to the last token, our approach enhances the reliability of model outputs without requiring additional training or external data. Evaluations on datasets including TruthfulQA, FACTOR, and HaluEval highlight the superior performance of our method in mitigating hallucinations, underscoring the necessity of ensuring accurate information flow in web-based applications. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2410_20340 |
| institution | arXiv |
| publishDate | 2024 |
| record_format | arxiv |
| spellingShingle | Maintaining Informative Coherence: Migrating Hallucinations in Large Language Models via Absorbing Markov Chains Wu, Jiemin Lai, Songning Xiao, Ruiqiang Xue, Tianlang Yang, Jiayu Yue, Yutao Computation and Language Artificial Intelligence Large Language Models (LLMs) are powerful tools for text generation, translation, and summarization, but they often suffer from hallucinations-instances where they fail to maintain the fidelity and coherence of contextual information during decoding, sometimes overlooking critical details due to their sampling strategies and inherent biases from training data and fine-tuning discrepancies. These hallucinations can propagate through the web, affecting the trustworthiness of information disseminated online. To address this issue, we propose a novel decoding strategy that leverages absorbing Markov chains to quantify the significance of contextual information and measure the extent of information loss during generation. By considering all possible paths from the first to the last token, our approach enhances the reliability of model outputs without requiring additional training or external data. Evaluations on datasets including TruthfulQA, FACTOR, and HaluEval highlight the superior performance of our method in mitigating hallucinations, underscoring the necessity of ensuring accurate information flow in web-based applications. |
| title | Maintaining Informative Coherence: Migrating Hallucinations in Large Language Models via Absorbing Markov Chains |
| topic | Computation and Language Artificial Intelligence |
| url | https://arxiv.org/abs/2410.20340 |