While reinforcement learning (RL) has been proven essential for tuning large language models (LLMs), it can lead to reward over-optimization (ROO). Existing approaches address ROO by adding KL regularization, requiring computationally expensive hyperparameter tuning. Additionally, KL regularization focuses solely on regularizing the language policy, neglecting a potential source of regularization: the reward function itself. Inspired by demonstration-guided RL, we here introduce the Reward Calibration from Demonstration (RCfD), which leverages human demonstrations and a reward model to recalibrate the reward objective. Formally, given a prompt, the RCfD objective minimizes the distance between the demonstrations’ and LLM’s rewards rather than directly maximizing the reward function. This objective shift avoids incentivizing the LLM to exploit the reward model and promotes more natural and diverse language generation.We show the effectiveness of RCfD in three RL language tasks, where it achieves comparable performance to carefully tuned baselines while mitigating ROO.
论文检索
输入标题、作者或关键词，从 9,460 篇学术成果中精准定位
CIF-Bench: A Chinese Instruction-Following Benchmark for Evaluating the Generalizability of Large Language Models
PDF ↗The advancement of large language models (LLMs) has enhanced the ability to generalize across a wide range of unseen natural language processing (NLP) tasks through instruction-following.Yet, their effectiveness often diminishes in low-resource languages like Chinese, exacerbated by biased evaluations from data leakage, casting doubt on their true generalizability to new linguistic territories. In response, we introduce the Chinese Instruction-Following Benchmark (**CIF-Bench**), designed to evaluate the zero-shot generalizability of LLMs to the Chinese language. CIF-Bench comprises 150 tasks and 15,000 input-output pairs, developed by native speakers to test complex reasoning and Chinese cultural nuances across 20 categories. To mitigate data contamination, we release only half of the dataset publicly, with the remainder kept private, and introduce diversified instructions to minimize score variance, totaling 45,000 data instances.Our evaluation of 28 selected LLMs reveals a noticeable performance gap, with the best model scoring only 52.9%, highlighting the limitations of LLMs in less familiar language and task contexts.This work not only uncovers the current limitations of LLMs in handling Chinese language tasks but also sets a new standard for future LLM generalizability research, pushing towards the development of more adaptable, culturally informed, and linguistically diverse models.
In the past year, MultiModal Large Language Models (MM-LLMs) have undergone substantial advancements, augmenting off-the-shelf LLMs to support MM inputs or outputs via cost-effective training strategies. The resulting models not only preserve the inherent reasoning and decision-making capabilities of LLMs but also empower a diverse range of MM tasks. In this paper, we provide a comprehensive survey aimed at facilitating further research of MM-LLMs. Initially, we outline general design formulations for model architecture and training pipeline. Subsequently, we introduce a taxonomy encompassing 126 MM-LLMs, each characterized by its specific formulations. Furthermore, we review the performance of selected MM-LLMs on mainstream benchmarks and summarize key training recipes to enhance the potency of MM-LLMs. Finally, we explore promising directions for MM-LLMs while concurrently maintaining a [real-time tracking website](https://mm-llms.github.io/) for the latest developments in the field. We hope that this survey contributes to the ongoing advancement of the MM-LLMs domain.
Large language models (LLMs) can generate long-form and coherent text, yet they often hallucinate facts, which undermines their reliability. To mitigate this issue, inference-time methods steer LLM representations toward the “truthful directions” previously learned for truth elicitation. However, applying these truthful directions with the same intensity fails to generalize across different query contexts. We propose LITO, a Learnable Intervention method for Truthfulness Optimization that automatically identifies the optimal intervention intensity tailored to each specific context. LITO explores a sequence of model generations based on increasing levels of intervention intensities. It selects the most accurate response or refuses to answer when the predictions are highly uncertain. Experiments on multiple LLMs and question-answering datasets demonstrate that LITO improves truthfulness while preserving task accuracy. The adaptive nature of LITO counters the limitations of one-size-fits-all intervention methods, maximizing truthfulness by reflecting the model’s internal knowledge only when it is confident. Our code is available at https://github.com/launchnlp/LITO.
Stance detection is a challenging task that aims to identify public opinion from social media platforms with respect to specific targets. Previous work on stance detection largely focused on pure texts. In this paper, we study multi-modal stance detection for tweets consisting of texts and images, which are prevalent in today’s fast-growing social media platforms where people often post multi-modal messages. To this end, we create five new multi-modal stance detection datasets of different domains based on Twitter, in which each example consists of a text and an image. In addition, we propose a simple yet effective Targeted Multi-modal Prompt Tuning framework (TMPT), where target information is leveraged to learn multi-modal stance features from textual and visual modalities. Experimental results on our five benchmark datasets show that the proposed TMPT achieves state-of-the-art performance in multi-modal stance detection.
In non-autoregressive translation (NAT), directed acyclic Transformers (DAT) have demonstrated their ability to achieve comparable performance to the autoregressive Transformers.In this paper, we first show that DAT is essentially a fully connected left-to-right Hidden Markov Model (HMM), with the source and target sequences being observations and the token positions being latent states.Even though generative models like HMM do not suffer from label bias in traditional task settings (e.g., sequence labeling), we argue here that the left-to-right HMM in NAT may still encounter this issue due to the missing observations at the inference stage.To combat label bias, we propose two constrained HMMs: 1) Adaptive Window HMM, which explicitly balances the number of outgoing transitions at different states; 2) Bi-directional HMM, i.e., a combination of left-to-right and right-to-left HMMs, whose uni-directional components can implicitly regularize each other’s biases via shared parameters.Experimental results on WMT’14 EnDe and WMT’17 ZhEn demonstrate that our methods can achieve better or comparable performance to the original DAT using various decoding methods.We also demonstrate that our methods effectively reduce the impact of label bias.
Beyond Text: Leveraging Multi-Task Learning and Cognitive Appraisal Theory for Post-Purchase Intention Analysis
PDF ↗Supervised machine-learning models for predicting user behavior offer a challenging classification problem with lower average prediction performance scores than other text classification tasks. This study evaluates multi-task learning frameworks grounded in Cognitive Appraisal Theory to predict user behavior as a function of users’ self-expression and psychological attributes. Our experiments show that users’ language and traits improve predictions above and beyond models predicting only from text. Our findings highlight the importance of integrating psychological constructs into NLP to enhance the understanding and prediction of user actions. We close with a discussion of the implications for future applications of large language models for computational psychology.
Many recent language models (LMs) are capable of in-context learning (ICL), manifested in the LMs’ ability to perform a new task solely from natural-language instruction. Previous work curating in-context learners assumes that ICL emerges from a vast over-parametrization or the scale of multi-task training. However, recent theoretical work attributes the ICL ability to concept-dependent training data and creates functional in-context learners even in small-scale, synthetic settings.In this work, we practically explore this newly identified axis of ICL quality. We propose Concept-aware Training (CoAT), a framework for constructing training scenarios that make it beneficial for the LM to learn to utilize the analogical reasoning concepts from demonstrations. We find that by using CoAT, pre-trained transformers can learn to better utilise new latent concepts from demonstrations and that such ability makes ICL more robust to the functional deficiencies of the previous models. Finally, we show that concept-aware in-context learners are much more effective in in-context learning a majority of unseen tasks compared to traditional instruction tuning, and fare comparably also to previous in-context learners trained in large-scale multitask learning requiring magnitudes of more training data.
Large Language Models are Few-Shot Training Example Generators: A Case Study in Fallacy Recognition
PDF ↗Recognizing fallacies is crucial for ensuring the quality and validity of arguments across various domains. However, computational fallacy recognition faces challenges due to the diverse genres, domains, and types of fallacies found in datasets. This leads to a highly multi-class, and even multi-label, setup with substantial class imbalance. In this study, we aim to enhance existing models for fallacy recognition by incorporating additional context and by leveraging large language models to generate synthetic data, thus increasing the representation of the infrequent classes. We experiment with GPT3.5 to generate synthetic examples and we examine the impact of prompt settings for this. Moreover, we explore zero-shot and few-shot scenarios to evaluate the effectiveness of using the generated examples for training smaller models within a unified fallacy recognition framework. Furthermore, we analyze the overlap between the synthetic data and existing fallacy datasets. Finally, we investigate the usefulness of providing supplementary context for detecting fallacy types that need such context, e.g., diversion fallacies. Our evaluation results demonstrate consistent improvements across fallacy types, datasets, and generators. The code and the synthetic datasets are all publicly available.
Argumentative Essay Generation (AEG) is a challenging task in computational argumentation, where detailed logical reasoning and effective rhetorical skills are essential.Previous methods on argument generation typically involve planning prior to generation.However, the planning strategies in these methods overlook the exploration of the logical reasoning process.Inspired by argument structure-related theories, we propose an argumentative planning strategy for prompting large language models (LLMs) to generate high-quality essays.This strategy comprises two stages: (1) Sketch planning, which creates a rough outline of the essay, and (2) Dialectical planning, which refines the outline through critical self-reflection.Such a planning strategy enables LLMs to write argumentative essays that are more logical, diverse, and persuasive.Furthermore, due to the scarcity of existing AEG datasets, we construct three new datasets.These datasets are from two domains: exam essays and news editorials, covering both Chinese and English.Automatic and manual evaluation on four datasets show that our method can generate more dialectical and persuasive essays with higher diversity compared to several strong baselines.
As humans, we consistently interact with our peers and receive feedback in the form of natural language. This language feedback allows us to maintain appropriate behavior, and rectify potential errors. The question arises naturally: can we use language feedback to align large language models (LLMs)? In contrast to previous research that aligns LLMs with scalar rewards, we present the first systematic exploration of alignment through the lens of language feedback (i.e., judgment). We start with an in-depth investigation of potential methods that can be adapted for aligning LLMs with judgments, revealing that these methods cannot fully capitalize on judgments. To facilitate more effective utilization of judgments, we propose a novel framework, Contrastive Unlikelihood Training (CUT), that allows for fine-grained inappropriate content detection and correction based on judgments. Our results show that, with merely 1317 off-the-shelf judgment data, CUT can beat the 175B DaVinci003 and surpass the best baseline by 50.84 points on AlpacaEval using LLaMA2-13b. CUT can also align LLMs in an iterative fashion using up-to-date model-specific judgments, improving performance from 81.09 to 91.68 points on AlpacaEval using LLaMA2-chat-13b. Further analysis suggests that judgments hold greater potential in LLM alignment than rewards.
Automated audio captioning (AAC) aims to generate descriptions based on audio input, attracting exploration of emerging audio language models (ALMs). However, current evaluation metrics only provide a single score to assess the overall quality of captions without characterizing the nuanced difference by systematically going through an evaluation checklist. To this end, we propose the explainable and multi-factor audio captioning evaluation (X-ACE) paradigm. X-ACE identifies four main factors that constitute the majority of audio features, specifically sound event, source, attribute and relation. To assess a given caption from an ALM, it is firstly transformed into an audio graph, where each node denotes an entity in the caption and corresponds to a factor. On the one hand, graph matching is conducted from part to whole for a holistic assessment. On the other hand, the nodes contained within each factor are aggregated to measure the factor-level performance. The pros and cons of an ALM can be explicitly and clearly demonstrated through X-ACE, pointing out the direction for further improvements. Experiments show that X-ACE exhibits better correlation with human perception and can detect mismatches sensitively.
An important requirement for the reliable deployment of pre-trained large language models (LLMs) is the well-calibrated quantification of the uncertainty in their outputs. While the likelihood of predicting the next token is a practical surrogate of the data uncertainty learned during training, model uncertainty is challenging to estimate, i.e., due to lack of knowledge acquired during training. Prior efforts to quantify uncertainty of neural networks require specific architectures or (re-)training strategies, which are impractical to apply to LLMs with several billion parameters, or for black-box models where the architecture and parameters are not available. In this paper, we propose Bayesian Prompts Ensembles (BayesPE), a novel approach to effectively obtain well-calibrated uncertainty for the output of pre-trained LLMs. BayesPE computes output probabilities through a weighted ensemble of different, but semantically equivalent, task instruction prompts. The relative weights of the different prompts in the ensemble are estimated through approximate Bayesian variational inference over a small labeled validation set. We demonstrate that BayesPE approximates a Bayesian input layer for the LLM, providing a lower bound on the expected model error. In our extensive experiments, we show that BayesPE achieves significantly superior uncertainty calibration compared to several baselines over a range of natural language classification tasks, both in zero- and few-shot settings.
Neural Machine Translation (NMT) encounters challenges when translating in new domains and low-resource languages. To address these issues, researchers have proposed methods to integrate additional knowledge into NMT, such as translation memories (TMs). However, finding TMs that closely match the input sentence remains challenging, particularly in specific domains. On the other hand, monolingual data is widely accessible in most languages, and back-translation is seen as a promising approach for utilizing target language data. Nevertheless, it still necessitates additional training. In this paper, we introduce Pseudo-kNN-MT, a variant of k-nearest neighbor machine translation (kNN-MT) that utilizes target language data by constructing a pseudo datastore. Furthermore, we investigate the utility of large language models (LLMs) for the kNN component. Experimental results demonstrate that our approach exhibits strong domain adaptation capability in both high-resource and low-resource machine translation. Notably, LLMs are found to be beneficial for robust NMT systems.
Increasing the number of parameters in large language models (LLMs) usually improves performance in downstream tasks but raises compute and memory costs, making deployment difficult in resource-limited settings. Quantization techniques, which reduce the bits needed for model weights or activations with minimal performance loss, have become popular due to the rise of LLMs. However, most quantization studies use pre-trained LLMs, and the impact of quantization on instruction-tuned LLMs and the relationship between perplexity and benchmark performance of quantized LLMs are not well understood. Evaluation of quantized LLMs is often limited to language modeling and a few classification tasks, leaving their performance on other benchmarks unclear. To address these gaps, we propose a structured evaluation framework consisting of three critical dimensions: (1) knowledge & capacity, (2) alignment, and (3) efficiency, and conduct extensive experiments across ten diverse benchmarks. Our experimental results indicate that LLMs with 4-bit quantization can retain performance comparable to their non-quantized counterparts, and perplexity can serve as a proxy metric for quantized LLMs on most benchmarks. Furthermore, quantized LLMs with larger parameter scales can outperform smaller LLMs. Despite the memory savings achieved through quantization, it can also slow down the inference speed of LLMs. Consequently, substantial engineering efforts and hardware support are imperative to achieve a balanced optimization of decoding speed and memory consumption in the context of quantized LLMs.
STSPL-SSC: Semi-Supervised Few-Shot Short Text Clustering with Semantic text similarity Optimized Pseudo-Labels
PDF ↗This study introduces the Semantic Textual Similarity Pseudo-Label Semi-Supervised Clustering (STSPL-SSC) framework. The STSPL-SSC framework is designed to tackle the prevalent issue of scarce labeled data by combining a Semantic Textual Similarity Pseudo-Label Generation process with a Robust Contrastive Learning module. The process begins with employing k-means clustering on embeddings for initial pseudo-Label allocation. Then we use a Semantic Text Similarity-enhanced module to supervise the secondary clustering of pseudo-labels using labeled data to better align with the real clustering centers. Subsequently, an Adaptive Optimal Transport (AOT) approach fine-tunes the pseudo-labels. Finally, a Robust Contrastive Learning module is employed to foster the learning of classification and instance-level distinctions, aiding clusters to better separate. Experiments conducted on multiple real-world datasets demonstrate that with just one label per class, clustering performance can be significantly improved, outperforming state-of-the-art models with an increase of 1-6% in both accuracy and normalized mutual information, approaching the results of fully-labeled classification.
Multilingual Large Language Models (LLMs) achieve remarkable levels of zero-shot cross-lingual transfer performance. We speculate that this is predicated on their ability to align languages without explicit supervision from parallel sentences. While representations of translationally equivalent sentences in different languages are known to be similar after convergence, however, it remains unclear how such cross-lingual alignment emerges during pre-training of LLMs. Our study leverages intrinsic probing techniques, which identify which subsets of neurons encode linguistic features, to correlate the degree of cross-lingual neuron overlap with the zero-shot cross-lingual transfer performance for a given model. In particular, we rely on checkpoints of BLOOM, a multilingual autoregressive LLM, across different training steps and model scales. We observe a high correlation between neuron overlap and downstream performance, which supports our hypothesis on the conditions leading to effective cross-lingual transfer. Interestingly, we also detect a degradation of both implicit alignment and multilingual abilities in certain phases of the pre-training process, providing new insights into the multilingual pretraining dynamics.
Hierarchical text classification aims at categorizing texts into a multi-tiered tree-structured hierarchy of labels. Existing methods pay more attention to capture hierarchy-aware text feature by exploiting explicit parent-child relationships, while interactions between peer labels are rarely taken into account, resulting in severe label confusion within each layer. In this work, we propose a novel Dual Prompt Tuning (DPT) method, which emphasizes identifying discrimination among peer labels by performing contrastive learning on each hierarchical layer. We design an innovative hand-crafted prompt containing slots for both positive and negative label predictions to cooperate with contrastive learning. In addition, we introduce a label hierarchy self-sensing auxiliary task to ensure cross-layer label consistency. Extensive experiments demonstrate that DPT achieves significant improvements and outperforms the current state-of-the-art methods on BGC and RCV1-V2 benchmark datasets.
While supervised fine-tuning (SFT) has been a straightforward approach for tailoring the output of foundation large language model (LLM) to specific preferences, concerns have been raised about the depth of this alignment, with some critiques suggesting it is merely “superficial”. We critically examine this hypothesis within the scope of cross-lingual generation tasks, proposing that the effectiveness of SFT may be constrained by its reliance on prior tokens to guide cross-lingual generation. Based on this crucial insight, and in response to the challenges posed by the costly and limited availability of non-English data for SFT, we introduce a novel training-free alignment method named PreTTY, which employs minimal task-related prior tokens to bridge the foundation LLM and the SFT LLM, achieving comparable performance without training. Experiments on machine translation and part-of-speech tagging across seven languages demonstrate the efficacy of PreTTY in cross-lingual settings. Remarkably, by initiating the decoding process with only one or two prior tokens, foundation LLMs can attain up to 98% of the performance metrics of their SFT counterparts. This method presents a cost-effective alternative to traditional SFT and advances the democratization of multilingual LLMs.
Large Language Models (LLMs), often show strong performance on English tasks, while exhibiting limitations on other languages. What is an LLM’s multilingual capability when it is trained only on certain languages? The underlying mechanism remains unclear. This study endeavors to examine the multilingual capability of LLMs from the vocabulary sharing perspective by conducting an exhaustive analysis across 101 languages. Through the investigation of the performance gap before and after embedding fine-tuning, we discovered four distinct quadrants. By delving into each quadrant we provide actionable and efficient guidelines for tuning these languages. Extensive experiments reveal that existing LLMs possess multilingual capabilities that surpass our expectations, and we can significantly improve the multilingual performance of LLMs based on these attributes of each quadrant .