Thailand

Social bias is shaped by the accumulation of social perceptions towards targets across various demographic identities. To fully understand such social bias in large language models (LLMs), it is essential to consider the composite of social perceptions from diverse perspectives among identities. Previous studies have either evaluated biases in LLMs by indirectly assessing the presence of sentiments towards demographic identities in the generated text or measuring the degree of alignment with given stereotypes. These methods have limitations in directly quantifying social biases at the level of distinct perspectives among identities. In this paper, we aim to investigate how social perceptions from various viewpoints contribute to the development of social bias in LLMs. To this end, we propose a novel strategy to intuitively quantify these social perceptions and suggest metrics that can evaluate the social biases within LLMs by aggregating diverse social perceptions. The experimental results show the quantitative demonstration of the social attitude in LLMs by examining social perception. The analysis we conducted shows that our proposed metrics capture the multi-dimensional aspects of social bias, enabling a fine-grained and comprehensive investigation of bias in LLMs.

ACL 2024

Ask LLMs Directly, &quot;What shapes your bias?&quot;: Measuring Social Bias in Large Language Models

persona effect

bias measurements

llm evaluation

social perception

social bias

### Welcome!
The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. Our Virtual Poster Sessions will take place online Thursday, August 22, 2024.

You are required to register for this event. **Please register [here](https://2024.aclweb.org/registration). **

If you have already registered, please check your inbox for an email from Underline granting you access to ACL 2024 content.

Please register!

The 62nd Annual Meeting of the Association for Computational Linguistics (ACL 2024) will take place in Bangkok, Thailand from August 11th to 16th, 2024. More information will be announced soon.

Ask LLMs Directly, "What shapes your bias?": Measuring Social Bias in Large Language Models

poster

Temporal Knowledge Graph (TKG) forecasting aims to predict future facts based on given histories. Most recent graph-based models excel at capturing structural information within TKGs but lack semantic comprehension abilities. Nowadays, with the surge of LLMs, the LLM-based TKG prediction model has emerged. However, the existing LLM-based model exhibits three shortcomings: (1) It only focuses on the first-order history for prediction while ignoring high-order historical information, resulting in the provided information for LLMs being extremely limited. (2) LLMs struggle with optimal reasoning performance under heavy historical information loads. (3) For TKG prediction, the temporal reasoning capability of LLM alone is limited. To address the first two challenges, we propose Chain-of-History (CoH) reasoning which explores high-order histories step-by-step, achieving effective utilization of high-order historical information for LLMs on TKG prediction. To address the third issue, we design CoH as a plug-and-play module to enhance the performance of graph-based models for TKG prediction. Extensive experiments on three datasets and backbones demonstrate the effectiveness of CoH.

Chain-of-History Reasoning for Temporal Knowledge Graph Forecasting

Making LLMs speak for different, especially minority groups of people, and generate statements supporting their diverse or even controversial perspectives is critical to creating an inclusive environment. However, existing LLMs lack sufficient controllability to the stance of their generated content, which often contains inconsistent, neutral, or biased statements. In this paper, we improve the controllability of LLMs in generating statements supporting an argument the user defined in the prompt. We find that multi-round debates between two LLMs with opposite stances generate higher-quality and more salient statements for each, which are important training data to improve the controllability of LLMs. Motivated by this, we develop a novel debate & tuning (“DEBATUNE”) pipeline finetuning LLMs to generate the statements obtained via debate. To examine DEBATUNE, we curate the largest dataset of debate topics so far, which covers 710 controversial topics and corresponding arguments for each topic. Evaluations by the GPT-4 judge with a novel controversy controllability metric show that LLMs’ capability of generating diverse perspectives is significantly improved by DEBATUNE. Moreover, such controllability can be generalized to unseen topics, generating high-quality statements supporting controversial arguments.

Can LLMs Speak For Diverse People? Tuning LLMs via Debate to Generate Controllable Controversial Statements

Detecting implicit hate speech that is not directly hateful remains a challenge. Recent research has attempted to detect implicit hate speech by applying contrastive learning to pre-trained language models such as BERT and RoBERTa, but the proposed models still do not have a significant advantage over cross-entropy loss-based learning. We found that contrastive learning based on randomly sampled batch data does not encourage the model to learn hard negative samples. In this work, we propose Label-aware Hard Negative sampling strategies (LAHN) that encourage the model to learn detailed features from hard negative samples, instead of naive negative samples in random batch, using momentum-integrated contrastive learning. LAHN outperforms the existing models for implicit hate speech detection both in- and cross-datasets. The code is available at https://github.com/Hanyang-HCC-Lab/LAHN

Label-aware Hard Negative Sampling Strategies with Momentum Contrastive Learning for Implicit Hate Speech Detection

In this paper, we aim to evaluate multi-agent systems against complex dependencies, including spatial, causal, and temporal constraints. First, we construct a new benchmark, named VillagerBench, within the Minecraft environment. VillagerBench comprises diverse tasks crafted to test various aspects of multi-agent collaboration, from workload distribution to dynamic adaptation and synchronized task execution. Second, we introduce a Directed Acyclic Graph Multi-Agent Framework (VillagerAgent) to resolve complex inter-agent dependencies and enhance collaborative efficiency. This solution incorporates a task decomposer that creates a directed acyclic graph (DAG) for structured task management, an agent controller for task distribution, and a state manager for tracking environmental and agent data.Our empirical evaluation on VillagerBench demonstrates that VillagerAgentoutperforms the existing AgentVerse model, reducing hallucinations and improving task decomposition efficacy. The results underscore VillagerAgent's potential in advancing multi-agent collaboration, offering a scalable and generalizable solution in dynamic environments. Source code is open-source on GitHub.

VillagerAgent: A Graph-Based Multi-Agent Framework for Coordinating Complex Task Dependencies in Minecraft

To process contexts with unlimited length using Large Language Models (LLMs), recent studies explore hierarchically managing the long text. Only several text fragments are taken from the external memory and passed into the temporary working memory, i.e., LLM's context window. However, existing approaches isolatedly handle the text fragments without considering their structural connections, thereby suffering limited capability on texts with intensive inter-relations, e.g., coherent stories and code repositories. This work attempts to resolve this by exploiting the fragment-level relations in external memory. First, we formulate the fragment-level relations and present several instantiations for different text types. Next, we introduce a relation-aware fragment assessment criteria upon previous independent fragment assessment. Finally, we present the fragment-connected Hierarchical Memory based LLM. We validate the benefits of involving these relations on long story understanding, repository-level code generation, and long-term chatting.

FragRel: Exploiting Fragment-level Relations in the External Memory of Large Language Models

The gap between the trepidation of program reliability and the expense of repairs underscore the indispensability for Automated Program Repair (APR). APR is instrumental in transforming vulnerable programs into more robust ones, bolstering program reliability while simultaneously diminishing the financial burden of manual repairs. Commercial-scale language models (LM) have taken APR to unprecedented levels. However, due to the limitations of model capabilities by parameters, a one-step substantial modification may not achieve the desired effect for models with parameters less than 100B. Moreover, humans interact with the LLM through explicit prompts, which hinders the LLM from receiving feedback from compiler and test cases to automatically optimize its repair policies. Explicit prompts from humans not only increase additional manpower costs, but also pose potential misunderstandings between human's intent and LMs.
Based on the above considerations, we are exploring how to ensure small-scale LM still outperform through process supervision and feedback. We start by constructing a dataset named CodeNet4Repair, replete with multiple repair records, which supervises the fine-tuning of a foundational mode. Building upon the encouraging outcomes of reinforcement learning, we develop a reward model that serves as a critic, providing feedback for the fine-tuned LM's action, progressively optimizing its policy. During inference, we require the LM to generate solutions iteratively until the repair effect no longer improves or hits the maximum step limit. The experimental results show that this process-based feedback not only outperforms larger outcome-based generation methods, but also nearly matches the performance of closed-source commercial large-scale LMs.

RePair: Automated Program Repair with Process-based Feedback

Through reading the documentation in the context, tool-using language models can dynamically extend their capability using external tools. The cost is that we have to input lengthy documentation every time the model needs to use the tool, occupying the input window as well as slowing down the decoding process.

Given the progress in general-purpose compression, soft context compression is a suitable approach to alleviate the problem. However, when compressing tool documentation, existing methods suffer from the weaknesses of key information loss (specifically, tool/parameter name errors) and difficulty in adjusting the length of compressed sequences based on documentation lengths.

To address these problems, we propose two strategies for compressing tool documentation into concise and precise summary sequences for tool-using language models. 1) Selective compression strategy mitigates key information loss by deliberately retaining key information as raw text tokens. 2) Block compression strategy involves dividing tool documentation into short chunks and then employing a fixed-length compression model to achieve variable-length compression. This strategy facilitates the flexible adjustment of the compression ratio.

Results on API-Bank and APIBench show that our approach reaches a performance comparable to the upper-bound baseline under up to 16x compression ratio.

Concise and Precise Context Compression for Tool-Using Language Models

Crafting an appealing heading is crucial for attracting readers and marketing work or products. A popular way is to summarize the main idea with a refined description and a memorable acronym. However, there lacks a systematic study and a formal benchmark including datasets and metrics. Motivated by this absence, we introduce LOgogram, a novel benchmark comprising 6,653 paper abstracts with corresponding descriptions and acronyms. To measure the quality of heading generation, we propose a set of evaluation metrics from three aspects: summarization, neology, and algorithm. Additionally, we explore three strategies for heading generation(generation ordering, tokenization of acronyms, and framework design) under various prevalent learning paradigms(supervised fine-tuning, in-context learning with Large Language Models(LLMs), and reinforcement learning) on our benchmark. Our experimental results indicate the difficulty in identifying a practice that excels across all summarization, neologistic, and algorithmic aspects.

Unveiling the Art of Heading Design: A Harmonious Blend of Summarization, Neology, and Algorithm

Relation extraction (RE) is a well-known NLP application often treated as a sentence or document-level task. However, a handful of recent efforts explore it across documents or in the cross-document setting (CrossDocRE). This is distinct from the single document case because different documents often focus on disparate themes, while text within a document tends to have a single goal.
Current CrossDocRE efforts do not consider domain knowledge, which are often assumed to be known to the reader when documents are authored. Here, we propose a novel approach, KXDocRE, that embed domain knowledge of entities with input text for cross-document RE. Our proposed framework has three main benefits over baselines: 1) it incorporates domain knowledge of entities along with documents' text; 2) it offers interpretability by producing explanatory text for predicted relations between entities 3) it improves performance over the prior methods. Code and models are available at~\url{https://github.com/kracr/cross-doc-relation-extraction}.

Knowledge-Driven Cross-Document Relation Extraction

In this work, we utilize Large Language Models (LLMs) for a novel use case: constructing Performance Predictors (PP) that estimate the performance of specific deep neural network architectures on downstream tasks. We create PP prompts for LLMs, comprising (i) role descriptions, (ii) instructions for the LLM, (iii) hyperparameter definitions, and (iv) demonstrations presenting sample architectures with efficiency metrics and `training from scratch' performance. In machine translation (MT) tasks, GPT-4 with our PP prompts (LLM-PP) achieves a SoTA mean absolute error and a slight degradation in rank correlation coefficient compared to baseline predictors. Additionally, we demonstrate that predictions from LLM-PP can be distilled to a compact regression model (LLM-Distill-PP), which surprisingly retains much of the performance of LLM-PP. This presents a cost-effective alternative for resource-intensive performance estimation. Specifically, for Neural Architecture Search (NAS), we introduce a Hybrid-Search algorithm (HS-NAS) employing LLM-Distill-PP for the initial search stages and reverting to the baseline predictor later. HS-NAS performs similarly to SoTA NAS, reducing search hours by approximately 50%, and in some cases, improving latency, GFLOPs, and model size. The code can be found at: https://github.com/UBC-NLP/llmas.

Downloads

Next from ACL 2024

Chain-of-History Reasoning for Temporal Knowledge Graph Forecasting

Similar lecture

Ask LLMs Directly, “What shapes your bias?”: Measuring Social Bias in Large Language Models