Aligning large language models (LLMs) with human preferences is crucial for enhancing their utility in terms of helpfulness, truthfulness, safety, harmlessness, and interestingness. Existing methods for achieving this alignment often involve employing reinforcement learning from human feedback (RLHF) to fine-tune LLMs based on human labels assessing the relative quality of model responses. Nevertheless, RLHF is susceptible to instability during fine-tuning and presents challenges in implementation. Drawing inspiration from the emerging field of representation engineering (RepE), this study aims to identify relevant representations for high-level human preferences embedded in patterns of activity within an LLM and achieve precise control of model behavior by transforming its representations. This novel approach, denoted as Representation Alignment from Human Feedback (RAHF), proves to be effective, computationally efficient, and easy to implement. Extensive experiments demonstrate the efficacy of RAHF in not only capturing but also manipulating representations to align with a broad spectrum of human preferences or values, rather than being confined to a singular concept or function (e.g. honesty or bias). RAHF’s versatility in accommodating diverse human preferences shows its potential for advancing LLM performance.

Parameter Efficient Fine-Tuning (PEFT) has gained significant attention for its ability to achieve competitive results while updating only a small subset of trainable parameters. Despite the promising performance of current PEFT methods, they present challenges in hyperparameter selection, such as determining the rank of LoRA or Adapter, or specifying the length of soft prompts. In addressing these challenges, we propose a novel approach to fine-tuning neural models, termed Representation EDiting (RED), which scales and biases the representation produced at each layer. RED substantially reduces the number of trainable parameters by a factor of 25,700 compared to full parameter fine-tuning, and by a factor of 32 compared to LoRA. Remarkably, RED achieves comparable or superior results to full parameter fine-tuning and other PEFT methods. Extensive experiments were conducted across models of varying architectures and scales, including RoBERTa, GPT-2, T5, and Llama-2, and the results demonstrate the efficiency and efficacy of RED, positioning it as a promising PEFT approach for large neural models.

The goal of multilingual modelling is to generate multilingual text representations for various downstream tasks in different languages. However, some state-of-the-art pre-trained multilingual models perform poorly on many low-resource languages due to the lack of representation space and model capacity. To alleviate this issue, we propose a Multilingual model Enhanced with Visual Text Representations (MEVTR), which complements textual representations and extends the multilingual representation space with visual text representations. First, the visual encoder focuses on the glyphs and structure of the text to obtain visual text representations, and the textual encoder obtains textual representations. Then, multilingual representations are enhanced by aligning and fusing visual text representations and textual representations. Moreover, we propose similarity constraint, a self-supervised task to prompt the visual encoder to focus on more additional information. Prefix alignment and multi-head bilinear module are designed to acquire an improved integration effect of visual text representations and textual representations. Experimental results indicate that MEVTR benefits from visual text representations and achieves significant performance gains in downstream tasks. In particular, in the zero-shot cross-lingual transfer task, MEVTR achieves results that outperform the state-of-the-art adapter-based framework without the target language adapter.

Recently, much research in psychology has benefited from the advances in machine learning techniques. Some recent studies showed that it is possible to build automated scoring models for children’s mindreading. These models were trained on a set of manually-labeled question-response pairs, which were collected by asking children to answer one or two questions after a short story is told or a video clip is played. However, existing models did not take the features of the stories and video clips into account when scoring, which obviously will reduce the accuracy of the scoring models. Furthermore, considering that different psychological tests may contain the same questions, this approach cannot be extended to other related psychological test datasets. In this study, we proposed a multi-modal learning framework to leverage the features extracted from the stories and videos related to the questions being asked during the children’s mindreading evaluation. Experimental results show that the scores produced by the proposed models agree well with those graded by human experts, highlighting the potential of the proposed network architecture for practical automated children’s mindreading scoring systems.

Large Language Models (LLMs) have made remarkable advancements in the field of natural language generation. However, the propensity of LLMs to generate inaccurate or non-factual content, termed “hallucinations”, remains a significant challenge. Current hallucination detection methods often necessitate the retrieval of great numbers of relevant evidence, thereby increasing response times. We introduce a unique framework that leverages statistical decision theory and Bayesian sequential analysis to optimize the trade-off between costs and benefits during the hallucination detection process. This approach does not require a predetermined number of observations. Instead, the analysis proceeds in a sequential manner, enabling an expeditious decision towards “belief” or “disbelief” through a stop-or-continue strategy. Extensive experiments reveal that this novel framework surpasses existing methods in both efficiency and precision of hallucination detection. Furthermore, it requires fewer retrieval steps on average, thus decreasing response times.

In this paper, we explore a new approach based on discourse analysis for the task of intent segmentation. Our target texts are user queries from a real-world chatbot. Our results show the feasibility of our approach with an F1-score of 82.97 points, and some advantages and disadvantages compared to two machine learning baselines: BERT and LSTM+CRF.