Large language models (LLMs) have shown surprisingly good performance in multilingual neural machine translation (MNMT) even if not being trained explicitly for translation. Yet, they still struggle with translating low-resource languages. As supported by our experiments, a bilingual dictionary between the source and the target language could help. Motivated by the fact that multilingual training effectively improves cross-lingual performance, we show that a chained multilingual dictionary with words expressed in more languages can provide more information to better enhance the LLM translation. To this end, we present a novel framework, CoD, Chain-of-Dictionary Prompting, which augments LLMs with prior knowledge with the chains of multilingual dictionaries for a subset of input words to elicit translation abilities for LLMs. Experiments indicate that ChatGPT and InstructGPT still have room for improvement in translating many language pairs. And CoD elicits large gains by up to 13x chrF++ points for MNMT (3.08 to 42.63 for English to Serbian written in Cyrillic script) on FLORES-200 full devtest set. We demonstrate the importance of chaining the multilingual dictionaries, as well as the superiority of CoD to few-shot in-context learning for low-resource languages. Using CoD helps ChatGPT to obviously surpass the SOTA translator NLLB 3.3B.

As the typical retraining paradigm is unacceptably time- and resource-consuming, researchers are turning to model editing to find an effective way that supports both consecutive and batch scenarios to edit the model behavior directly. Despite all these practical expectations, existing model editing methods fail to realize all of them. Furthermore, the memory demands for such sequential model editing approaches tend to be prohibitive, frequently necessitating an external memory that grows incrementally over time. To cope with these challenges, we propose CoachHooK, a model editing method that simultaneously supports sequential and batch editing. CoachHooK is memory-friendly as it only needs a small amount of it to store several hook layers whose size remains unchanged over time. Experimental results demonstrate the superiority of our method over other batch-supportive model editing methods under both single-round and consecutive batch editing scenarios. Extensive analyses of CoachHooK have been conducted to verify the stability of our method over a number of consecutive steps.

Despite the fact that multilingual agreement (MA) has shown its importance for multilingual neural machine translation (MNMT), current methodologies in the field have two shortages: (i) require parallel data between multiple language pairs, which is not always realistic and (ii) optimize the agreement in an ambiguous direction, which hampers the translation performance. We present Bidirectional Multilingual Agreement via Switched Back-translation (BMA-SBT), a novel and universal multilingual agreement framework for fine-tuning pre-trained MNMT models, which (i) exempts the need for aforementioned parallel data by using a novel method called switched BT that creates synthetic text written in another source language using the translation target and (ii) optimizes the agreement bidirectionally with the Kullback-Leibler Divergence loss. Experiments indicate that BMA-SBT clearly improves the strong baselines on the task of MNMT with three benchmarks: TED Talks, News, and Europarl. In-depth analyzes indicate that BMA-SBT brings additive improvements to the conventional BT method.

We are currently in an era of fierce competition among various large language models (LLMs), continuously pushing the boundaries of benchmark performance. However, genuinely assessing the capabilities of these LLMs has become a challenging and critical issue due to potential data contamination. In this paper, we propose a novel and valuable method, Clean-Eval, which mitigates the issue of data contamination and evaluates the LLMs more cleanly. Clean-Eval employs a neural-based model to paraphrase and back-translate the contaminated data into a candidate set, generating expressions with the same meaning but in different surface forms. A semantic detector is then used to filter those generated low-quality samples to narrow down this candidate set. Candidates with moderate BLEURT scores against the original samples are selected as the final evaluation set. According to human assessment, this set is almost semantically equivalent to the original contamination set but expressed differently. We conduct experiments on 20 existing benchmarks across diverse tasks, and results demonstrate that Clean-Eval substantially restores the actual evaluation results on contaminated LLMs under both few-shot learning and fine-tuning scenarios.

While Large Language Models (LLMs) have demonstrated exceptional multitasking abilities, fine-tuning these models on downstream, domain-specific datasets is often necessary to yield superior performance on test sets compared to their counterparts without fine-tuning. However, the comprehensive effects of fine-tuning on the LLMs’ generalization ability are not fully understood.This paper delves into the differences between original, unmodified LLMs and their fine-tuned variants. Our primary investigation centers on whether fine-tuning affects the generalization ability intrinsic to LLMs. To elaborate on this, we conduct extensive experiments across five distinct language tasks on various datasets.Our main findings reveal that models fine-tuned on generation and classification tasks exhibit dissimilar behaviors in generalizing to different domains and tasks.Intriguingly, we observe that integrating the in-context learning strategy during fine-tuning on generation tasks can enhance the model’s generalization ability.Through this systematic investigation, we aim to contribute valuable insights into the evolving landscape of fine-tuning practices for LLMs.

Most research about natural language generation (NLG) relies on evaluation benchmarks with limited references for a sample, which may result in poor correlations with human judgements. The underlying reason is that one semantic meaning can actually be expressed in different forms, and the evaluation with a single or few references may not accurately reflect the quality of the model’s hypotheses. To address this issue, this paper presents a simple and effective method, named **Div-Ref**, to enhance existing evaluation benchmarks by enriching the number of references. We leverage large language models (LLMs) to diversify the expression of a single reference into multiple high-quality ones to cover the semantic space of the reference sentence as much as possible. We conduct comprehensive experiments to empirically demonstrate that diversifying the expression of reference can significantly enhance the correlation between automatic evaluation and human evaluation. This idea is compatible with recent LLM-based evaluation which can similarly derive advantages from incorporating multiple references. *We strongly encourage future generation benchmarks to include more references, even if they are generated by LLMs, which is once for all.* We release all the code and data at https://github.com/RUCAIBox/Div-Ref to facilitate research.

In the realm of emerging multitasking abilities of Large language models (LLMs), methodologies like prompt tuning enable low-cost adaptation to downstream tasks without retraining the model. However, automatic input pre-processing when LLMs are unavailable is currently under-studied. This paper proposes ReLLM (Rephrasing for LLMs), a method that automatically paraphrases input content for better output generations. ReLLM replaces low-frequency lexical items with their high-frequency counterparts. This substitution is particularly beneficial for low-resource language tasks that lack sufficient training data and resources. ReLLM is user-friendly and requires no additional LLM training. Experimental results in cross-lingual summarization, and natural language inference demonstrate the effectiveness of ReLLM.

In this paper, we study the generalization ability of large language models (LLMs) with respect to compositional instructions, which are instructions that can be decomposed into several sub-instructions. We argue that the ability to generalize from simple instructions to more intricate compositional instructions represents a key aspect of the out-of-distribution generalization for LLMs. Since there are no specialized datasets for studying this phenomenon, we first construct a dataset with the help of ChatGPT, guided by the self-instruct technique. Then, we fine-tune and evaluate LLMs on these datasets. Interestingly, our experimental results indicate that training LLMs on higher-order compositional instructions enhances their performance on lower-order ones, but the reverse does not hold true.

Curriculum Data Augmentation (CDA) improves neural models by presenting synthetic data with increasing difficulties from easy to hard. However, traditional CDA simply treats the ratio of word perturbation as the difficulty measure and goes through the curriculums only once. This paper presents PCC: Paraphrasing with Bottom-k Sampling and Cyclic Learning for Curriculum Data Augmentation, a novel CDA framework via paraphrasing, which exploits the textual paraphrase similarity as the curriculum difficulty measure. We propose a curriculum-aware paraphrase generation module composed of three units: a paraphrase candidate generator with bottom-k sampling, a filtering mechanism and a difficulty measure. We also propose a cyclic learning strategy that passes through the curriculums multiple times. The bottom-k sampling is proposed to generate super-hard instances for the later curriculums. Experimental results on few-shot text classification as well as dialogue generation indicate that PCC surpasses competitive baselines. Human evaluation and extensive case studies indicate that bottom-k sampling effectively generates super-hard instances, and PCC significantly improves the baseline dialogue agent.

Despite the success of multilingual sequence-to-sequence pre-training, most existing approaches rely on document-level monolingual corpora in many different languages, sentence-level bilingual corpora, and sometimes synthetic document-level bilingual corpora. This hampers the performance with cross-lingual document-level tasks such as document-level translation. Hence, we propose to mine and leverage document-level trilingual parallel corpora to improve sequence-to-sequence multilingual pre-training. We present Triangular Document-level Pre-training (TRIP) as the first in the field to accelerate the conventional monolingual and bilingual objectives into a trilingual objective with a novel method called Grafting. Experiments show that TRIP achieves several strong state-of-the-art (SOTA) scores on three multilingual document-level machine translation benchmarks and one cross-lingual abstractive summarization benchmark, including consistent improvements by up to 3.11 d-BLEU points and 8.9 ROUGE-L points.

MultiDoc2Dial presents an important challenge on modeling dialogues grounded with multiple documents. This paper proposes a pipeline system of “retrieve, re-rank, and generate”, where each component is individually optimized. This enables the passage re-ranker and response generator to fully exploit training with ground-truth data. Furthermore, we use a deep cross-encoder trained with localized hard negative passages from the retriever. For the response generator, we use grounding span prediction as an auxiliary task to be jointly trained with the main task of response generation. We also adopt a passage dropout and regularization technique to improve response generation performance. Experimental results indicate that the system clearly surpasses the competitive baseline and our team CPII-NLP ranked 1st among the public submissions on ALL four leaderboards based on the sum of F1, SacreBLEU, METEOR and RougeL scores.

Dialogue agents can leverage external textual knowledge to generate responses of a higher quality. To our best knowledge, most existing works on knowledge grounded dialogue settings assume that the user intention is always answerable. Unfortunately, this is impractical as there is no guarantee that the knowledge retrievers could always retrieve the desired knowledge. Therefore, this is crucial to incorporate fallback responses to respond to unanswerable contexts appropriately while responding to the answerable contexts in an informative manner. We propose a novel framework that automatically generates a control token with the generator to bias the succeeding response towards informativeness for answerable contexts and fallback for unanswerable contexts in an end-to-end manner. Since no existing knowledge grounded dialogue dataset considers this aim, we augment the existing dataset with unanswerable contexts to conduct our experiments. Automatic and human evaluation results indicate that naively incorporating fallback responses with controlled text generation still hurts informativeness for answerable context. In contrast, our proposed framework effectively mitigates this problem while still appropriately presenting fallback responses to unanswerable contexts. Such a framework also reduces the extra burden of the additional classifier and the overheads introduced in the previous works, which operates in a pipeline manner.

Incorporating personas information allows diverse and engaging responses in dialogue response generation. Unfortunately, prior works have primarily focused on self personas and have overlooked the value of partner personas. Moreover, in practical applications, the availability of the gold partner personas is often not the case. This paper attempts to tackle these issues by offering a novel framework that leverages automatic partner personas generation to enhance the succeeding dialogue response generation. Our framework employs reinforcement learning with a dedicatedly designed critic network for reward judgement. Experimental results from automatic and human evaluations indicate that our framework is capable of generating relevant, interesting, coherent and informative partner personas, even compared to the ground truth partner personas. This enhances the succeeding dialogue response generation, which surpasses our competitive baselines that condition on the ground truth partner personas.