Given the prompt "Rome is in", can we steer a language model to flip its prediction of an incorrect token "France" to a correct token "Italy" by only multiplying a few relevant activation...
详细信息
Implicit Personalization (IP) is a phenomenon of language models inferring a user's background from the implicit cues in the input prompts and tailoring the response based on this inference. While previous work ha...
Recent studies have shown that many naturallanguage understanding and reasoning datasets contain statistical cues that can be exploited by NLP models, resulting in an overestimation of their capabilities. Existing me...
详细信息
ISBN:
(纸本)9798891760615
Recent studies have shown that many naturallanguage understanding and reasoning datasets contain statistical cues that can be exploited by NLP models, resulting in an overestimation of their capabilities. Existing methods, such as "hypothesis-only" tests and CheckList, are limited in identifying these cues and evaluating model weaknesses. We introduce ICQ (I-See-Cue), a lightweight, general statistical profiling framework that automatically identifies potential biases in multiple-choice NLU datasets without requiring additional test cases. ICQ assesses the extent to which models exploit these biases through black-box testing, addressing the limitations of current methods. In this work, we conduct a comprehensive evaluation of statistical biases in 10 popular NLU datasets and 4 models, confirming prior findings, revealing new insights, and offering an online demonstration system to encourage users to assess their own datasets and models. Furthermore, we present a case study on investigating ChatGPT's bias, providing valuable recommendations for practical applications.
The recent emergence of Medical Large Vision language Models (Med-LVLMs) has enhanced medical diagnosis. However, current Med-LVLMs frequently encounter factual issues, often generating responses that do not align wit...
详细信息
ISBN:
(纸本)9798891761643
The recent emergence of Medical Large Vision language Models (Med-LVLMs) has enhanced medical diagnosis. However, current Med-LVLMs frequently encounter factual issues, often generating responses that do not align with established medical facts. Retrieval-Augmented Generation (RAG), which utilizes external knowledge, can improve the factual accuracy of these models but introduces two major challenges. First, limited retrieved contexts might not cover all necessary information, while excessive retrieval can introduce irrelevant and inaccurate references, interfering with the model's generation. Second, in cases where the model originally responds correctly, applying RAG can lead to an over-reliance on retrieved contexts, resulting in incorrect answers. To address these issues, we propose RULE, which consists of two components. First, we introduce a provably effective strategy for controlling factuality risk through the calibrated selection of the number of retrieved contexts. Second, based on samples where over-reliance on retrieved contexts led to errors, we curate a preference dataset to fine-tune the model, balancing its dependence on inherent knowledge and retrieved contexts for generation. We demonstrate the effectiveness of RULE on medical VQA and report generation tasks across three datasets, achieving an average improvement of 47.4% in factual accuracy. We publicly release our benchmark and code in https: //***/richard-peng- xia/RULE.
Large language models (LLMs) appear to bias their survey answers toward certain values. Nonetheless, some argue that LLMs are too inconsistent to simulate particular values. Are they? To answer, we first define value ...
详细信息
This paper presents a way of enhancing the reliability of Large Multi-modal Models (LMMs) in addressing hallucination, where the models generate cross-modal inconsistent responses. Without additional training, we prop...
详细信息
Existing works examining Vision-language Models (VLMs) for social biases predominantly focus on a limited set of documented bias associations, such as gender↔profession or race↔crime. This narrow scope often overlooks...
详细信息
Code switching (CS) is a very common phenomenon in written and spoken communication but one that is handled poorly by many naturallanguageprocessing (NLP) applications. Looking to the application of building CS corp...
详细信息
ISBN:
(纸本)9798891760882
Code switching (CS) is a very common phenomenon in written and spoken communication but one that is handled poorly by many naturallanguageprocessing (NLP) applications. Looking to the application of building CS corpora, we explore CS language identification (LID) for corpus building. We make the task more realistic by scaling it to more languages and considering models with simpler architectures for faster inference. We also reformulate the task as a sentence-level multi-label tagging problem to make it more tractable. Having defined the task, we investigate three reasonable models for this task and define metrics which better reflect desired performance. We present empirical evidence that no current approach is adequate and finally provide recommendations for future work in this area.
With an increase in complexity and severity, it is becoming harder to identify and mitigate vulnerabilities. Although traditional tools remain useful, machine learning models are being adopted to expand efforts. To he...
详细信息
With an increase in complexity and severity, it is becoming harder to identify and mitigate vulnerabilities. Although traditional tools remain useful, machine learning models are being adopted to expand efforts. To help explore methods of vulnerability detection, we present an empirical study on the effectiveness of text-based machine learning models by utilizing 344 open-source projects, 2,182 vulnerabilities and 38 vulnerability types. With the availability of vulnerabilities being presented in forms such as code snippets, we construct a methodology based on extracted source code functions and create equal pairings. We conduct experiments using seven machine learning models, five naturallanguageprocessing techniques and three data processingmethods. First, we present results based on full context function pairings. Next, we introduce condensed functions and conduct a statistical analysis to determine if there is a significant difference between the models, techniques, or methods. Based on these results, we answer research questions regarding model prediction for testing within and across projects and vulnerability types. Our results show that condensed functions with fewer features may achieve greater prediction results when testing within rather than across. Overall, we conclude that text-based machine learning models are not effective in detecting vulnerabilities within or across projects and vulnerability types.
Large language Models (LLMs) are increasingly ubiquitous, yet their ability to retain and reason about temporal information remains limited, hindering their application in real-world scenarios where understanding the ...
详细信息
暂无评论