Permissive Information-Flow Analysis for Large Language Models

• URL: http://arxiv.org/abs/2410.03055v1
• Date: Fri, 4 Oct 2024 00:25:43 GMT
• Title: Permissive Information-Flow Analysis for Large Language Models
• Authors: Shoaib Ahmed Siddiqui, Radhika Gaonkar, Boris Köpf, David Krueger, Andrew Paverd, Ahmed Salem, Shruti Tople, Lukas Wutschitz, Menglin Xia, Santiago Zanella-Béguelin,
• Abstract summary: Large Language Models (LLMs) are rapidly becoming commodity components of larger software systems. This poses natural security and privacy problems: poisoned data retrieved from one component can change the model's behavior and compromise the entire system. We propose a novel, more permissive approach to propagate information flow labels through LLM queries.
• Score: 21.563132267220073
• License: http://creativecommons.org/licenses/by-nc-sa/4.0/
• Abstract: Large Language Models (LLMs) are rapidly becoming commodity components of larger software systems. This poses natural security and privacy problems: poisoned data retrieved from one component can change the model's behavior and compromise the entire system, including coercing the model to spread confidential data to untrusted components. One promising approach is to tackle this problem at the system level via dynamic information flow (aka taint) tracking. Unfortunately, the traditional approach of propagating the most restrictive input label to the output is too conservative for applications where LLMs operate on inputs retrieved from diverse sources. In this paper, we propose a novel, more permissive approach to propagate information flow labels through LLM queries. The key idea behind our approach is to propagate only the labels of the samples that were influential in generating the model output and to eliminate the labels of unnecessary input. We implement and investigate the effectiveness of two variations of this approach, based on (i) prompt-based retrieval augmentation, and (ii) a $k$-nearest-neighbors language model. We compare these with the baseline of an introspection-based influence estimator that directly asks the language model to predict the output label. The results obtained highlight the superiority of our prompt-based label propagator, which improves the label in more than 85% of the cases in an LLM agent setting. These findings underscore the practicality of permissive label propagation for retrieval augmentation.

Related papers

• Zero-to-Strong Generalization: Eliciting Strong Capabilities of Large Language Models Iteratively without Gold Labels [75.77877889764073]
  Large Language Models (LLMs) have demonstrated remarkable performance through supervised fine-tuning or in-context learning using gold labels. This study explores whether solely utilizing unlabeled data can elicit strong model capabilities. We propose a new paradigm termed zero-to-strong generalization.
  arXiv  Detail & Related papers  (2024-09-19T02:59:44Z)
• Are you still on track!? Catching LLM Task Drift with Activations [55.75645403965326]
  Task drift allows attackers to exfiltrate data or influence the LLM's output for other users. We show that a simple linear classifier can detect drift with near-perfect ROC AUC on an out-of-distribution test set. We observe that this approach generalizes surprisingly well to unseen task domains, such as prompt injections, jailbreaks, and malicious instructions.
  arXiv  Detail & Related papers  (2024-06-02T16:53:21Z)
• TnT-LLM: Text Mining at Scale with Large Language Models [24.731544646232962]
  Large Language Models (LLMs) automate the process of end-to-end label generation and assignment with minimal human effort. We show that TnT-LLM generates more accurate and relevant label when compared against state-of-the-art baselines. We also share our practical experiences and insights on the challenges and opportunities of using LLMs for large-scale text mining in real-world applications.
  arXiv  Detail & Related papers  (2024-03-18T18:45:28Z)
• Measuring Distributional Shifts in Text: The Advantage of Language Model-Based Embeddings [11.393822909537796]
  An essential part of monitoring machine learning models in production is measuring input and output data drift. Recent advancements in large language models (LLMs) indicate their effectiveness in capturing semantic relationships. We propose a clustering-based algorithm for measuring distributional shifts in text data by exploiting such embeddings.
  arXiv  Detail & Related papers  (2023-12-04T20:46:48Z)
• Improving Input-label Mapping with Demonstration Replay for In-context Learning [67.57288926736923]
  In-context learning (ICL) is an emerging capability of large autoregressive language models. We propose a novel ICL method called Sliding Causal Attention (RdSca) We show that our method significantly improves the input-label mapping in ICL demonstrations.
  arXiv  Detail & Related papers  (2023-10-30T14:29:41Z)
• Substituting Data Annotation with Balanced Updates and Collective Loss in Multi-label Text Classification [19.592985329023733]
  Multi-label text classification (MLTC) is the task of assigning multiple labels to a given text. We study the MLTC problem in annotation-free and scarce-annotation settings in which the magnitude of available supervision signals is linear to the number of labels. Our method follows three steps, (1) mapping input text into a set of preliminary label likelihoods by natural language inference using a pre-trained language model, (2) calculating a signed label dependency graph by label descriptions, and (3) updating the preliminary label likelihoods with message passing along the label dependency graph.
  arXiv  Detail & Related papers  (2023-09-24T04:12:52Z)
• From Quantity to Quality: Boosting LLM Performance with Self-Guided Data Selection for Instruction Tuning [52.257422715393574]
  We introduce a self-guided methodology for Large Language Models (LLMs) to autonomously discern and select cherry samples from open-source datasets. Our key innovation, the Instruction-Following Difficulty (IFD) metric, emerges as a pivotal metric to identify discrepancies between a model's expected responses and its intrinsic generation capability.
  arXiv  Detail & Related papers  (2023-08-23T09:45:29Z)
• CELDA: Leveraging Black-box Language Model as Enhanced Classifier without Labels [14.285609493077965]
  Clustering-enhanced Linear Discriminative Analysis, a novel approach that improves the text classification accuracy with a very weak-supervision signal. Our framework draws a precise decision boundary without accessing weights or gradients of the LM model or data labels.
  arXiv  Detail & Related papers  (2023-06-05T08:35:31Z)
• Label-Retrieval-Augmented Diffusion Models for Learning from Noisy Labels [61.97359362447732]
  Learning from noisy labels is an important and long-standing problem in machine learning for real applications. In this paper, we reformulate the label-noise problem from a generative-model perspective. Our model achieves new state-of-the-art (SOTA) results on all the standard real-world benchmark datasets.
  arXiv  Detail & Related papers  (2023-05-31T03:01:36Z)
• Deep Partial Multi-Label Learning with Graph Disambiguation [27.908565535292723]
  We propose a novel deep Partial multi-Label model with grAph-disambIguatioN (PLAIN) Specifically, we introduce the instance-level and label-level similarities to recover label confidences. At each training epoch, labels are propagated on the instance and label graphs to produce relatively accurate pseudo-labels.
  arXiv  Detail & Related papers  (2023-05-10T04:02:08Z)
• Guiding Large Language Models via Directional Stimulus Prompting [114.84930073977672]
  We introduce Directional Stimulus Prompting, a novel framework for guiding black-box large language models (LLMs) toward specific desired outputs. Instead of directly adjusting LLMs, our method employs a small tunable policy model to generate an auxiliary directional stimulus prompt for each input instance.
  arXiv  Detail & Related papers  (2023-02-22T17:44:15Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.