Shammie
Abstract:LLMs often adopt an assertive language style also when making false claims. Such ``overconfident hallucinations'' mislead users and erode trust. Achieving the ability to express in language the actual degree of uncertainty around a claim is therefore of great importance. We find that ``verbal uncertainty'' is governed by a single linear feature in the representation space of LLMs, and show that this has only moderate correlation with the actual ``semantic uncertainty'' of the model. We apply this insight and show that (1) the mismatch between semantic and verbal uncertainty is a better predictor of hallucinations than semantic uncertainty alone and (2) we can intervene on verbal uncertainty at inference time and reduce hallucinations on short-form answers, achieving an average relative reduction of 32%.
Abstract:Large language models (LLMs) trained on massive multilingual datasets hint at the formation of interlingual constructs--a shared subspace in the representation space. However, evidence regarding this phenomenon is mixed, leaving it unclear whether these models truly develop unified interlingual representations, or present a partially aligned constructs. We explore 31 diverse languages varying on their resource-levels, typologies, and geographical regions; and find that multilingual LLMs exhibit inconsistent cross-lingual alignments. To address this, we propose an interlingual representation framework identifying both the shared interlingual semantic subspace and fragmented components, existed due to representational limitations. We introduce Interlingual Local Overlap (ILO) score to quantify interlingual alignment by comparing the local neighborhood structures of high-dimensional representations. We utilize ILO to investigate the impact of single-language fine-tuning on the interlingual representations in multilingual LLMs. Our results indicate that training exclusively on a single language disrupts the alignment in early layers, while freezing these layers preserves the alignment of interlingual representations, leading to improved cross-lingual generalization. These results validate our framework and metric for evaluating interlingual representation, and further underscore that interlingual alignment is crucial for scalable multilingual learning.
Abstract:Large Language Models often generate factually incorrect but plausible outputs, known as hallucinations. We identify a more insidious phenomenon, LLM delusion, defined as high belief hallucinations, incorrect outputs with abnormally high confidence, making them harder to detect and mitigate. Unlike ordinary hallucinations, delusions persist with low uncertainty, posing significant challenges to model reliability. Through empirical analysis across different model families and sizes on several Question Answering tasks, we show that delusions are prevalent and distinct from hallucinations. LLMs exhibit lower honesty with delusions, which are harder to override via finetuning or self reflection. We link delusion formation with training dynamics and dataset noise and explore mitigation strategies such as retrieval augmented generation and multi agent debating to mitigate delusions. By systematically investigating the nature, prevalence, and mitigation of LLM delusions, our study provides insights into the underlying causes of this phenomenon and outlines future directions for improving model reliability.
Abstract:Large language models (LLMs) frequently hallucinate due to misaligned self-awareness, generating erroneous outputs when addressing queries beyond their knowledge boundaries. While existing approaches mitigate hallucinations via uncertainty estimation or query rejection, they suffer from computational inefficiency or sacrificed helpfulness. To address these issues, we propose the Explicit Knowledge Boundary Modeling (EKBM) framework, integrating fast and slow reasoning systems to harmonize reliability and usability. The framework first employs a fast-thinking model to generate confidence-labeled responses, enabling immediate use of high-confidence outputs. For uncertain predictions, a slow refinement model conducts targeted reasoning to improve accuracy. To align model behavior with our proposed object, we propose a hybrid training pipeline, enhancing self-awareness without degrading task performance. Evaluations on dialogue state tracking tasks demonstrate that EKBM achieves superior model reliability over uncertainty-based baselines. Further analysis reveals that refinement substantially boosts accuracy while maintaining low computational overhead. Our work establishes a scalable paradigm for advancing LLM reliability and balancing accuracy and practical utility in error-sensitive applications.
Abstract:The first International AI Safety Report comprehensively synthesizes the current evidence on the capabilities, risks, and safety of advanced AI systems. The report was mandated by the nations attending the AI Safety Summit in Bletchley, UK. Thirty nations, the UN, the OECD, and the EU each nominated a representative to the report's Expert Advisory Panel. A total of 100 AI experts contributed, representing diverse perspectives and disciplines. Led by the report's Chair, these independent experts collectively had full discretion over the report's content.
Abstract:The hallucination problem of Large Language Models (LLMs) significantly limits their reliability and trustworthiness. Humans have a self-awareness process that allows us to recognize what we don't know when faced with queries. Inspired by this, our paper investigates whether LLMs can estimate their own hallucination risk before response generation. We analyze the internal mechanisms of LLMs broadly both in terms of training data sources and across 15 diverse Natural Language Generation (NLG) tasks, spanning over 700 datasets. Our empirical analysis reveals two key insights: (1) LLM internal states indicate whether they have seen the query in training data or not; and (2) LLM internal states show they are likely to hallucinate or not regarding the query. Our study explores particular neurons, activation layers, and tokens that play a crucial role in the LLM perception of uncertainty and hallucination risk. By a probing estimator, we leverage LLM self-assessment, achieving an average hallucination estimation accuracy of 84.32\% at run time.
Abstract:The capability to reason from text is crucial for real-world NLP applications. Real-world scenarios often involve incomplete or evolving data. In response, individuals update their beliefs and understandings accordingly. However, most existing evaluations assume that language models (LMs) operate with consistent information. We introduce Belief-R, a new dataset designed to test LMs' belief revision ability when presented with new evidence. Inspired by how humans suppress prior inferences, this task assesses LMs within the newly proposed delta reasoning ($\Delta R$) framework. Belief-R features sequences of premises designed to simulate scenarios where additional information could necessitate prior conclusions drawn by LMs. We evaluate $\sim$30 LMs across diverse prompting strategies and found that LMs generally struggle to appropriately revise their beliefs in response to new information. Further, models adept at updating often underperformed in scenarios without necessary updates, highlighting a critical trade-off. These insights underscore the importance of improving LMs' adaptiveness to changing information, a step toward more reliable AI systems.
Abstract:We introduce a formal information-theoretic framework for image captioning by regarding it as a representation learning task. Our framework defines three key objectives: task sufficiency, minimal redundancy, and human interpretability. Building upon this foundation, we propose a novel Pyramid of Captions (PoCa) method, which constructs caption pyramids by generating localized captions for zoomed-in image patches and integrating them with global caption information using large language models. This approach leverages intuition that the detailed examination of local patches can reduce error risks and address inaccuracies in global captions, either by correcting the hallucination or adding missing details. Based on our theoretical framework, we formalize this intuition and provide formal proof demonstrating the effectiveness of PoCa under certain assumptions. Empirical tests with various image captioning models and large language models show that PoCa consistently yields more informative and semantically aligned captions, maintaining brevity and interpretability.
Abstract:The widespread application of Large Language Models (LLMs) across various tasks and fields has necessitated the alignment of these models with human values and preferences. Given various approaches of human value alignment, ranging from Reinforcement Learning with Human Feedback (RLHF), to constitutional learning, etc. there is an urgent need to understand the scope and nature of human values injected into these models before their release. There is also a need for model alignment without a costly large scale human annotation effort. We propose UniVaR, a high-dimensional representation of human value distributions in LLMs, orthogonal to model architecture and training data. Trained from the value-relevant output of eight multilingual LLMs and tested on the output from four multilingual LLMs, namely LlaMA2, ChatGPT, JAIS and Yi, we show that UniVaR is a powerful tool to compare the distribution of human values embedded in different LLMs with different langauge sources. Through UniVaR, we explore how different LLMs prioritize various values in different languages and cultures, shedding light on the complex interplay between human values and language modeling.
Abstract:Large language models (LLMs) show remarkable human-like capability in various domains and languages. However, a notable quality gap arises in low-resource languages, e.g., Indonesian indigenous languages, rendering them ineffective and inefficient in such linguistic contexts. To bridge this quality gap, we introduce Cendol, a collection of Indonesian LLMs encompassing both decoder-only and encoder-decoder architectures across a range of model sizes. We highlight Cendol's effectiveness across a diverse array of tasks, attaining 20% improvement, and demonstrate its capability to generalize to unseen tasks and indigenous languages of Indonesia. Furthermore, Cendol models showcase improved human favorability despite their limitations in capturing indigenous knowledge and cultural values in Indonesia. In addition, we discuss the shortcomings of parameter-efficient tunings, such as LoRA, for language adaptation. Alternatively, we propose the usage of vocabulary adaptation to enhance efficiency. Lastly, we evaluate the safety of Cendol and showcase that safety in pre-training in one language such as English is transferable to low-resource languages, such as Indonesian, even without RLHF and safety fine-tuning.