相关论文: Surgical, Cheap, and Flexible: Mitigating False Re…

Adversarial learning can learn fairer and less biased models of language than standard methods. However, current adversarial techniques only partially mitigate model bias, added to which their training procedures are often unstable. In this…

The application of machine learning based decision making systems in safety critical areas requires reliable high certainty predictions. Reject options are a common way of ensuring a sufficiently high certainty of predictions made by the…

Refusal on harmful prompts is a key safety behaviour in instruction-tuned large language models (LLMs), yet the internal causes of this behaviour remain poorly understood. We study two public instruction-tuned models, Gemma-2-2B-IT and…

SafePredict is a novel meta-algorithm that works with any base prediction algorithm for online data to guarantee an arbitrarily chosen correctness rate, $1-\epsilon$, by allowing refusals. Allowing refusals means that the meta-algorithm may…

We study how to perform unlearning, i.e. forgetting undesirable misbehaviors, on large language models (LLMs). We show at least three scenarios of aligning LLMs with human preferences can benefit from unlearning: (1) removing harmful…

Machine learning has advanced dramatically, narrowing the accuracy gap to humans in multimodal tasks like visual question answering (VQA). However, while humans can say "I don't know" when they are uncertain (i.e., abstain from answering a…

Large Language Models (LLMs) commonly rely on explicit refusal prefixes for safety, making them vulnerable to prefix injection attacks. We introduce HumorReject, a novel data-driven approach that reimagines LLM safety by decoupling it from…

Pretrained vision-and-language BERTs aim to learn representations that combine information from both modalities. We propose a diagnostic method based on cross-modal input ablation to assess the extent to which these models actually…

Large Language Models used in ChatGPT have traditionally been trained to learn a refusal boundary: depending on the user's intent, the model is taught to either fully comply or outright refuse. While this is a strong mitigation for…

Large Language Models (LLMs) often expend significant computational resources generating boilerplate responses, such as refusals, simple acknowledgements and casual greetings, which adds unnecessary cost and latency. To address this…

Safety-trained language models routinely refuse requests for help circumventing rules. But not all rules deserve compliance. When users ask for help evading rules imposed by an illegitimate authority, rules that are deeply unjust or absurd…

Instruction-following language models often show undesirable biases. These undesirable biases may be accelerated in the real-world usage of language models, where a wide range of instructions is used through zero-shot example prompting. To…

We propose an approach for preventing unsafe or otherwise low-quality large language model (LLM) outputs by leveraging the stochasticity of LLMs, an approach we call Repeated Checking with Regeneration (RCR). In this system, LLM checkers…

Large language models (LLMs) have transformed the way we access information. These models are often tuned to refuse to comply with requests that are considered harmful and to produce responses that better align with the preferences of those…

Inasmuch as the removal of refusal behavior from instruction-tuned language models by directional abliteration requires the extraction of refusal-mediating directions from the residual stream activation space, and inasmuch as the…

Language models can be persuaded to abandon factual knowledge. This vulnerability is central to AI safety, but its internal mechanism remains poorly understood. We uncover a compact causal mechanism for persuasion-induced factual errors. A…

When language models lack relevant knowledge for a given query, they frequently generate plausible responses that can be hallucinations, rather than admitting being agnostic about the answer. Retraining models to reward admitting ignorance…

Recently developed large language models have achieved remarkable success in generating fluent and coherent text. However, these models often tend to 'hallucinate' which critically hampers their reliability. In this work, we address this…

In order to be deployed safely, Large Language Models (LLMs) must be capable of dynamically adapting their behavior based on their level of knowledge and uncertainty associated with specific topics. This adaptive behavior, which we refer to…

Credit scoring models support loan approval decisions in the financial services industry. Lenders train these models on data from previously granted credit applications, where the borrowers' repayment behavior has been observed. This…