Large language models (LLMs) have revolutionized NLP research. Notably, in-context learning enables their use as evaluation metrics for natural language generation, making them particularly advantageous in low-resource scenarios and time-restricted applications. In this work, we introduce PrExMe, a large-scale Prompt Exploration for Metrics, where we evaluate more than 720 prompt templates for open-source LLM-based metrics on machine translation (MT) and summarization datasets, totalling over 6.6M evaluations. This extensive comparison (1) benchmarks recent open-source LLMs as metrics and (2) explores the stability and variability of different prompting strategies. We discover that, on the one hand, there are scenarios for which prompts are stable. For instance, some LLMs show idiosyncratic preferences and favor to grade generated texts with textual labels while others prefer to return numeric scores. On the other hand, the stability of prompts and model rankings can be susceptible to seemingly innocuous changes. For example, changing the requested output format from "0 to 100" to "-1 to +1" can strongly affect the rankings in our evaluation. Our study contributes to understanding the impact of different prompting approaches on LLM-based metrics for MT and summarization evaluation, highlighting the most stable prompting patterns and potential limitations.
Large language models (LLMs) have shown the emergent capability of in-context learning (ICL). One line of research has claimed that ICL is functionally equivalent to gradient descent, a type of error-driven learning mechanism. In this paper, we introduce a new way of diagnosing whether ICL is functionally performing error-driven learning. Our approach is based on the inverse frequency effect (IFE) -- a phenomenon in which an agent's behavior is influenced to a greater degree when presented with improbable examples as compared to more likely ones. The IFE has previously been identified in psycholinguistics where humans exhibit the IFE in the context of structural priming (the tendency for people to produce sentence structures they have encountered recently). In that context, the IFE has been used as evidence that human structural priming must involve error-driven learning mechanisms. In our experiments, we simulated structural priming with ICL and found that LLMs indeed display the IFE, with the effect being stronger in larger models. We conclude that at least in the case we studied, ICL is indeed a type of error-driven learning, supporting the hypothesis that an error signal is implicitly computed in the forward pass during ICL. Our results suggest that both humans and LLMs make use of error-driven processing mechanisms in on-line processing.
Vafa et al. (2024) introduced a transformer-based econometric model, CAREER, that predicts a worker's next job as a function of career history (an "occupation model"). CAREER was initially estimated ("pre-trained") using a large, unrepresentative resume dataset, which served as a "foundation model," and parameter estimation was continued ("fine-tuned") using data from a representative survey. CAREER had better predictive performance than benchmarks. This paper considers an alternative where the resume-based foundation model is replaced by a large language model (LLM). We convert tabular data from the survey into text files that resemble resumes and fine-tune the LLMs using these text files with the objective to predict the next token (word). The resulting fine-tuned LLM is used as an input to an occupation model. Its predictive performance surpasses all prior models. We demonstrate the value of fine-tuning and further show that by adding more career data from a different population, fine-tuning smaller LLMs surpasses the performance of fine-tuning larger models.
Within few-shot learning, in-context learning (ICL) has become a potential method for leveraging contextual information to improve model performance on small amounts of data or in resource-constrained environments where training models on large datasets is prohibitive. However, the quality of the selected sample in a few shots severely limits the usefulness of ICL. The primary goal of this paper is to enhance the performance of evaluation metrics for in-context learning by selecting high-quality samples in few-shot learning scenarios. We employ the chi-square test to identify high-quality samples and compare the results with those obtained using low-quality samples. Our findings demonstrate that utilizing high-quality samples leads to improved performance with respect to all evaluated metrics.
Badminton enjoys widespread popularity, and reports on matches generally include details such as player names, game scores, and ball types, providing audiences with a comprehensive view of the games. However, writing these reports can be a time-consuming task. This challenge led us to explore whether a Large Language Model (LLM) could automate the generation and evaluation of badminton reports. We introduce a novel framework named BADGE, designed for this purpose using LLM. Our method consists of two main phases: Report Generation and Report Evaluation. Initially, badminton-related data is processed by the LLM, which then generates a detailed report of the match. We tested different Input Data Types, In-Context Learning (ICL), and LLM, finding that GPT-4 performs best when using CSV data type and the Chain of Thought prompting. Following report generation, the LLM evaluates and scores the reports to assess their quality. Our comparisons between the scores evaluated by GPT-4 and human judges show a tendency to prefer GPT-4 generated reports. Since the application of LLM in badminton reporting remains largely unexplored, our research serves as a foundational step for future advancements in this area. Moreover, our method can be extended to other sports games, thereby enhancing sports promotion. For more details, please refer to https://github.com/AndyChiangSH/BADGE.
The extraction of lung lesion information from clinical and medical imaging reports is crucial for research on and clinical care of lung-related diseases. Large language models (LLMs) can be effective at interpreting unstructured text in reports, but they often hallucinate due to a lack of domain-specific knowledge, leading to reduced accuracy and posing challenges for use in clinical settings. To address this, we propose a novel framework that aligns generated internal knowledge with external knowledge through in-context learning (ICL). Our framework employs a retriever to identify relevant units of internal or external knowledge and a grader to evaluate the truthfulness and helpfulness of the retrieved internal-knowledge rules, to align and update the knowledge bases. Experiments with expert-curated test datasets demonstrate that this ICL approach can increase the F1 score for key fields (lesion size, margin and solidity) by an average of 12.9% over existing ICL methods.
In the Minecraft Collaborative Building Task, two players collaborate: an Architect (A) provides instructions to a Builder (B) to assemble a specified structure using 3D blocks. In this work, we investigate the use of large language models (LLMs) to predict the sequence of actions taken by the Builder. Leveraging LLMs' in-context learning abilities, we use few-shot prompting techniques, that significantly improve performance over baseline methods. Additionally, we present a detailed analysis of the gaps in performance for future work
Hierarchical text classification (HTC) is an important task with broad applications, while few-shot HTC has gained increasing interest recently. While in-context learning (ICL) with large language models (LLMs) has achieved significant success in few-shot learning, it is not as effective for HTC because of the expansive hierarchical label sets and extremely-ambiguous labels. In this work, we introduce the first ICL-based framework with LLM for few-shot HTC. We exploit a retrieval database to identify relevant demonstrations, and an iterative policy to manage multi-layer hierarchical labels. Particularly, we equip the retrieval database with HTC label-aware representations for the input texts, which is achieved by continual training on a pretrained language model with masked language modeling (MLM), layer-wise classification (CLS, specifically for HTC), and a novel divergent contrastive learning (DCL, mainly for adjacent semantically-similar labels) objective. Experimental results on three benchmark datasets demonstrate superior performance of our method, and we can achieve state-of-the-art results in few-shot HTC.
Low-resource languages, by its very definition, tend to be under represented in the pre-training corpora of Large Language Models. In this work, we investigate three low-resource cross-lingual approaches that enable an LLM adapt to tasks in previously unseen languages. Llama-2 is an LLM where Indic languages, among many other language families, contribute to less than $0.005\%$ of the total $2$ trillion token pre-training corpora. In this work, we experiment with the English-dominated Llama-2 for cross-lingual transfer to three Indic languages, Bengali, Hindi, and Tamil as target languages. We study three approaches for cross-lingual transfer, under ICL and fine-tuning. One, we find that adding additional supervisory signals via a dominant language in the LLM, leads to improvements, both under in-context learning and fine-tuning. Two, adapting the target languages to word reordering may be beneficial under ICL, but its impact diminishes with fine tuning. Finally, continued pre-training in one low-resource language can improve model performance for other related low-resource languages.
Decompilation transforms compiled code back into a high-level programming language for analysis when source code is unavailable. Previous work has primarily focused on enhancing decompilation performance by increasing the scale of model parameters or training data for pre-training. Based on the characteristics of the decompilation task, we propose two methods: (1) Without fine-tuning, the Self-Constructed Context Decompilation (sc$^2$dec) method recompiles the LLM's decompilation results to construct pairs for in-context learning, helping the model improve decompilation performance. (2) Fine-grained Alignment Enhancement (FAE), which meticulously aligns assembly code with source code at the statement level by leveraging debugging information, is employed during the fine-tuning phase to achieve further improvements in decompilation. By integrating these two methods, we achieved a Re-Executability performance improvement of approximately 3.90% on the Decompile-Eval benchmark, establishing a new state-of-the-art performance of 52.41%. The code, data, and models are available at https://github.com/AlongWY/sccdec.
Computational social science (CSS) practitioners often rely on human-labeled data to fine-tune supervised text classifiers. We assess the potential for researchers to augment or replace human-generated training data with surrogate training labels from generative large language models (LLMs). We introduce a recommended workflow and test this LLM application by replicating 14 classification tasks and measuring performance. We employ a novel corpus of English-language text classification data sets from recent CSS articles in high-impact journals. Because these data sets are stored in password-protected archives, our analyses are less prone to issues of contamination. For each task, we compare supervised classifiers fine-tuned using GPT-4 labels against classifiers fine-tuned with human annotations and against labels from GPT-4 and Mistral-7B with few-shot in-context learning. Our findings indicate that supervised classification models fine-tuned on LLM-generated labels perform comparably to models fine-tuned with labels from human annotators. Fine-tuning models using LLM-generated labels can be a fast, efficient and cost-effective method of building supervised text classifiers.
Tool learning aims to enhance and expand large language models' (LLMs) capabilities with external tools, which has gained significant attention recently. Current methods have shown that LLMs can effectively handle a certain amount of tools through in-context learning or fine-tuning. However, in real-world scenarios, the number of tools is typically extensive and irregularly updated, emphasizing the necessity for a dedicated tool retrieval component. Tool retrieval is nontrivial due to the following challenges: 1) complex user instructions and tool descriptions; 2) misalignment between tool retrieval and tool usage models. To address the above issues, we propose to enhance tool retrieval with iterative feedback from the large language model. Specifically, we prompt the tool usage model, i.e., the LLM, to provide feedback for the tool retriever model in multi-round, which could progressively improve the tool retriever's understanding of instructions and tools and reduce the gap between the two standalone components. We build a unified and comprehensive benchmark to evaluate tool retrieval models. The extensive experiments indicate that our proposed approach achieves advanced performance in both in-domain evaluation and out-of-domain evaluation.
Large language models (LLMs) possess extensive parametric knowledge, but this knowledge is difficult to update with new information because retraining is very expensive and infeasible for closed-source models. Knowledge editing (KE) has emerged as a viable solution for updating the knowledge of LLMs without compromising their overall performance. On-the-fly KE methods, inspired by in-context learning (ICL), have shown great promise and allow LLMs to be treated as black boxes. In the past, KE was primarily employed in English contexts, whereas the potential for cross-lingual KE in current English-centric LLMs has not been fully explored. To foster more research in this direction, we introduce the BMIKE-53 benchmark for evaluating cross-lingual KE on 53 diverse languages across three KE task types. We also propose a gradient-free KE method called Multilingual In-context Knowledge Editing (MIKE) and evaluate it on BMIKE-53. Our evaluation focuses on cross-lingual knowledge transfer in terms of reliability, generality, locality, and portability, offering valuable insights and a framework for future research in cross-lingual KE. Our code and data are publicly accessible via the anonymous repository at https://anonymous.4open.science/r/MIKE.
Large Language Models (LLMs) have demonstrated remarkable abilities, one of the most important being In-Context Learning (ICL). With ICL, LLMs can derive the underlying rule from a few demonstrations and provide answers that comply with the rule. Previous work hypothesized that the network creates a "task vector" in specific positions during ICL. Patching the "task vector" allows LLMs to achieve zero-shot performance similar to few-shot learning. However, we discover that such "task vectors" do not exist in tasks where the rule has to be defined through multiple demonstrations. Instead, the rule information provided by each demonstration is first transmitted to its answer position and forms its own rule vector. Importantly, all the rule vectors contribute to the output in a distributed manner. We further show that the rule vectors encode a high-level abstraction of rules extracted from the demonstrations. These results are further validated in a series of tasks that rely on rules dependent on multiple demonstrations. Our study provides novel insights into the mechanism underlying ICL in LLMs, demonstrating how ICL may be achieved through an information aggregation mechanism.
Efficient fine-tuning of large language models for task-specific applications is imperative, yet the vast number of parameters in these models makes their training increasingly challenging. Despite numerous proposals for effective methods, a substantial memory overhead remains for gradient computations during updates. \thm{Can we fine-tune a series of task-specific small models and transfer their knowledge directly to a much larger model without additional training?} In this paper, we explore weak-to-strong specialization using logit arithmetic, facilitating a direct answer to this question. Existing weak-to-strong methods often employ a static knowledge transfer ratio and a single small model for transferring complex knowledge, which leads to suboptimal performance. % To address this, To surmount these limitations, we propose a dynamic logit fusion approach that works with a series of task-specific small models, each specialized in a different task. This method adaptively allocates weights among these models at each decoding step, learning the weights through Kullback-Leibler divergence constrained optimization problems. We conduct extensive experiments across various benchmarks in both single-task and multi-task settings, achieving leading results. By transferring expertise from the 7B model to the 13B model, our method closes the performance gap by 96.4\% in single-task scenarios and by 86.3\% in multi-task scenarios compared to full fine-tuning of the 13B model. Notably, we achieve surpassing performance on unseen tasks. Moreover, we further demonstrate that our method can effortlessly integrate in-context learning for single tasks and task arithmetic for multi-task scenarios.
Recently, there has been increasing interest in applying large language models (LLMs) as zero-shot passage rankers. However, few studies have explored how to select appropriate in-context demonstrations for the passage ranking task, which is the focus of this paper. Previous studies mainly use LLM's feedback to train a retriever for demonstration selection. These studies apply the LLM to score each demonstration independently, which ignores the dependencies between demonstrations (especially important in ranking task), leading to inferior performance of top-$k$ retrieved demonstrations. To mitigate this issue, we introduce a demonstration reranker to rerank the retrieved demonstrations so that top-$k$ ranked ones are more suitable for ICL. However, generating training data for such reranker is quite challenging. On the one hand, different from demonstration retriever, the training samples of reranker need to incorporate demonstration dependencies. On the other hand, obtaining the gold ranking from the retrieved demonstrations is an NP-hard problem, which is hard to implement. To overcome these challenges, we propose a method to approximate the optimal demonstration list iteratively and utilize LLM to score demonstration lists of varying lengths. By doing so, the search space is greatly reduced and demonstration dependencies are considered. Based on these scored demonstration lists, we further design a list-pairwise training approach which compares a pair of lists that only differ in the last demonstration, to teach the reranker how to select the next demonstration given a previous sequence. In this paper, we propose a demonstration selection framework DemoRank for ranking task and conduct extensive experiments to prove its strong ability.
The path to interpreting a language model often proceeds via analysis of circuits -- sparse computational subgraphs of the model that capture specific aspects of its behavior. Recent work has automated the task of discovering circuits. Yet, these methods have practical limitations, as they rely either on inefficient search algorithms or inaccurate approximations. In this paper, we frame automated circuit discovery as an optimization problem and propose *Edge Pruning* as an effective and scalable solution. Edge Pruning leverages gradient-based pruning techniques, but instead of removing neurons or components, it prunes the \emph{edges} between components. Our method finds circuits in GPT-2 that use less than half the number of edges compared to circuits found by previous methods while being equally faithful to the full model predictions on standard circuit-finding tasks. Edge Pruning is efficient even with as many as 100K examples, outperforming previous methods in speed and producing substantially better circuits. It also perfectly recovers the ground-truth circuits in two models compiled with Tracr. Thanks to its efficiency, we scale Edge Pruning to CodeLlama-13B, a model over 100x the scale that prior methods operate on. We use this setting for a case study comparing the mechanisms behind instruction prompting and in-context learning. We find two circuits with more than 99.96% sparsity that match the performance of the full model and reveal that the mechanisms in the two settings overlap substantially. Our case study shows that Edge Pruning is a practical and scalable tool for interpretability and sheds light on behaviors that only emerge in large models.
Today, the detection of AI-generated content is receiving more and more attention. Our idea is to go beyond detection and try to recover the prompt used to generate a text. This paper, to the best of our knowledge, introduces the first investigation in this particular domain without a closed set of tasks. Our goal is to study if this approach is promising. We experiment with zero-shot and few-shot in-context learning but also with LoRA fine-tuning. After that, we evaluate the benefits of using a semi-synthetic dataset. For this first study, we limit ourselves to text generated by a single model. The results show that it is possible to recover the original prompt with a reasonable degree of accuracy.
This work investigates the in-context learning abilities of pretrained large language models (LLMs) when instructed to translate text from a low-resource language into a high-resource language as part of an automated machine translation pipeline. We conduct a set of experiments translating Southern Quechua to Spanish and examine the informativity of various types of context retrieved from a constrained database of digitized pedagogical materials (dictionaries and grammar lessons) and parallel corpora. Using both automatic and human evaluation of model output, we conduct ablation studies that manipulate (1) context type (morpheme translations, grammar descriptions, and corpus examples), (2) retrieval methods (automated vs. manual), and (3) model type. Our results suggest that even relatively small LLMs are capable of utilizing prompt context for zero-shot low-resource translation when provided a minimally sufficient amount of relevant linguistic information. However, the variable effects of context type, retrieval method, model type, and language-specific factors highlight the limitations of using even the best LLMs as translation systems for the majority of the world's 7,000+ languages and their speakers.
Objective: Radiotherapy treatment planning is a time-consuming and potentially subjective process that requires the iterative adjustment of model parameters to balance multiple conflicting objectives. Recent advancements in frontier Artificial Intelligence (AI) models offer promising avenues for addressing the challenges in planning and clinical decision-making. This study introduces GPT-RadPlan, an automated treatment planning framework that integrates radiation oncology knowledge with the reasoning capabilities of large multi-modal models, such as GPT-4Vision (GPT-4V) from OpenAI. Approach: Via in-context learning, we incorporate clinical requirements and a few (3 in our experiments) approved clinical plans with their optimization settings, enabling GPT-4V to acquire treatment planning domain knowledge. The resulting GPT-RadPlan system is integrated into our in-house inverse treatment planning system through an application programming interface (API). For a given patient, GPT-RadPlan acts as both plan evaluator and planner, first assessing dose distributions and dose-volume histograms (DVHs), and then providing textual feedback on how to improve the plan to match the physician's requirements. In this manner, GPT-RadPlan iteratively refines the plan by adjusting planning parameters, such as weights and dose objectives, based on its suggestions. Main results: The efficacy of the automated planning system is showcased across 17 prostate cancer and 13 head and neck cancer VMAT plans with prescribed doses of 70.2 Gy and 72 Gy, respectively, where we compared GPT-RadPlan results to clinical plans produced by human experts. In all cases, GPT-RadPlan either outperformed or matched the clinical plans, demonstrating superior target coverage and reducing organ-at-risk doses by 5 Gy on average (15 percent for prostate and 10-15 percent for head and neck).