This page is still being polished. If you have thoughts, please share them via the feedback form.
Data on this page is preliminary and may change. Please do not share or cite these figures publicly.
Changes to the model's learned parameters, architecture, or training process, including modifications to training data that affect what the model learns.
Also in AI System
Privacy leakage is a crucial risk of LLMs, since the powerful memorization and association capabilities of LLMs raise the risk of revealing private information within the training data. Researchers are devoted to designing privacypreserving frameworks in LLMs [226], [227], aiming to safeguard sensitive PII from possible disclosure during humanmachine conservation
Reasoning
Privacy-preserving frameworks modify training objectives and procedures to reduce memorization and prevent information leakage.
Private Data Interventions
The intervention can be accomplished by lexicon-based approaches [228] or trainable classifiers [229]–[231]. The lexicon-based approaches are usually based on pre-defined rules to recognize and cleanse sensitive PII entities. Alternatively, recent work tends to employ neural networks to automate the intervention process. For instance, the developers of GPT-4 have built automatic models to identify and remove the PII entities within the training data [2]. A number of evaluation studies [231], [232] demonstrated that the methods of data intervention like deduplication and text sanitization are able to effectively improve the safety of LLMs (e.g., GPT-3.5 and LLaMA-7B) in privacy.
1.1.1 Training DataPrivacy Enhanced Techniques
Differential privacy (DP) [233]–[235] is a type of randomized algorithm to protect a private dataset from privacy leakage. To preserve individual information memorized by the model, developers can train the model with a differential privacy guarantee to hide the difference between two neighboring datasets (only one element is different between the two datasets). The goal of DP algorithms is to leave an acceptable distance that makes the two datasets indistinguishable. Lots of efforts have developed DP techniques as the standard for protecting privacy in earlier transformer-based PLMs and LLMs [236]–[238].
1.1.2 Learning ObjectivesMitigation in Input Modules
Mitigating the threat posed by the input module presents a significant challenge for LLM developers due to the diversity of the harmful inputs and adversarial prompts [209], [210].
1.2.1 Guardrails & FilteringMitigation in Input Modules > Defensive Prompt Design
Directly modifying the input prompts is a viable approach to steer the behavior of the model and foster the generation of responsible outputs. This method integrates contextual information or constraints in the prompts to provide background knowledge and guidelines while generating the output [22].
1.2.1 Guardrails & FilteringMitigation in Input Modules > Malicious Prompt Detection
Different from the methods of designing defensive prompts to preprocess the input, the malicious prompt detection method aims to detect and filter out the harmful prompts through the input safeguard
1.2.1 Guardrails & FilteringMitigation in Language Models
This section delves into mitigating risks associated with models, encompassing privacy preservation, detoxification and debiasing, mitigation of hallucinations, and defenses against model attacks.
1.1 ModelMitigation in Language Models > Detoxifying and Debiasing
To reduce the toxicity and bias of LLMs, prior efforts mainly focus on enhancing the quality of training data and conducting safety training.
1.1 ModelMitigation in Language Models > Hallucination Mitigation
Hallucinations, one of the key challenges associated with LLMs, have received extensive studies.
1.2.9 OtherRisk Taxonomy, Mitigation, and Assessment Benchmarks of Large Language Model Systems
Cui, Tianyu; Wang, Yanling; Fu, Chuanpu; Xiao, Yong; Li, Sijia; Deng, Xinhao; Liu, Yunpeng; Zhang, Qinglin; Qiu, Ziyi; Li, Peiyang; Tan, Zhixing; Xiong, Junwu; Kong, Xinyu; Wen, Zujie; Xu, Ke; Li, Qi (2024)
Despite their impressive capabilities, large lan- guage models (LLMs) have been observed to generate responses that include inaccurate or fabricated information, a phenomenon com- monly known as “hallucination”. In this work, we propose a simple Induce-then-Contrast De- coding (ICD) strategy to alleviate hallucina- tions. We first construct a factually weak LLM by inducing hallucinations from the original LLMs. Then, we penalize these induced hallu- cinations during decoding to enhance the fac- tuality of the generated content. Concretely, we determine the final next-token predictions by amplifying the predictions from the orig- inal model and downplaying the induced un- truthful predictions via contrastive decoding. Experimental results on both discrimination- based and generation-based hallucination eval- uation benchmarks, such as TruthfulQA and FACTSCORE, demonstrate that our proposed ICD methods can effectively enhance the factu- ality of LLMs across various model sizes and families. For example, when equipped with ICD, Llama2-7B-Chat and Mistral-7B-Instruct achieve performance comparable to ChatGPT and GPT4 on TruthfulQA, respectively.
Build and Use Model
Training, fine-tuning, and integrating the AI model
Developer
Entity that creates, trains, or modifies the AI system
Manage
Prioritising, responding to, and mitigating AI risks
