Large language models trained for safety and harmlessness remain susceptible to adversarial misuse, as evidenced by the prevalence of "jailbreak" attacks on early releases of ChatGPT that elicit undesired behavior. Going beyond recognition of the issue, we investigate why such attacks succeed and how they can be created. We hypothesize two failure modes of safety training: competing objectives and mismatched generalization. Competing objectives arise when a model's capabilities and safety goals conflict, while mismatched generalization occurs when safety training fails to generalize to a domain for which capabilities exist. We use these failure modes to guide jailbreak design and then evaluate state-of-the-art models, including OpenAI's GPT-4 and Anthropic's Claude v1.3, against both existing and newly designed attacks. We find that vulnerabilities persist despite the extensive red-teaming and safety-training efforts behind these models. Notably, new attacks utilizing our failure modes succeed on every prompt in a collection of unsafe requests from the models' red-teaming evaluation sets and outperform existing ad hoc jailbreaks. Our analysis emphasizes the need for safety-capability parity -- that safety mechanisms should be as sophisticated as the underlying model -- and argues against the idea that scaling alone can resolve these safety failure modes.

Current natural language systems designed for multi-step claim validation typically operate in two phases: retrieve a set of relevant premise statements using heuristics (planning), then generate novel conclusions from those statements using a large language model (deduction). The planning step often requires expensive Transformer operations and does not scale to arbitrary numbers of premise statements. In this paper, we investigate whether an efficient planning heuristic is possible via embedding spaces compatible with deductive reasoning. Specifically, we evaluate whether embedding spaces exhibit a property we call deductive additivity: the sum of premise statement embeddings should be close to embeddings of conclusions based on those premises. We explore multiple sources of off-the-shelf dense embeddings in addition to fine-tuned embeddings from GPT3 and sparse embeddings from BM25. We study embedding models both intrinsically, evaluating whether the property of deductive additivity holds, and extrinsically, using them to assist planning in natural language proof generation. Lastly, we create a dataset, Single-Step Reasoning Contrast (SSRC), to further probe performance on various reasoning types. Our findings suggest that while standard embedding methods frequently embed conclusions near the sums of their premises, they fall short of being effective heuristics and lack the ability to model certain categories of reasoning.

Large language models (LLMs) have brought significant and transformative changes in human society. These models have demonstrated remarkable capabilities in natural language understanding and generation, leading to various advancements and impacts across several domains. We consider the in-context learning under two formulation for attention related regression in this work. Given matrices $A_1 \in \mathbb{R}^{n \times d}$, and $A_2 \in \mathbb{R}^{n \times d}$ and $B \in \mathbb{R}^{n \times n}$, the purpose is to solve some certain optimization problems: Normalized version $\min_{X} \| D(X)^{-1} \exp(A_1 X A_2^\top) - B \|_F^2$ and Rescaled version $\| \exp(A_1 X A_2^\top) - D(X) \cdot B \|_F^2$. Here $D(X) := \mathrm{diag}( \exp(A_1 X A_2^\top) {\bf 1}_n )$. Our regression problem shares similarities with previous studies on softmax-related regression. Prior research has extensively investigated regression techniques related to softmax regression: Normalized version $\| \langle \exp(Ax) , {\bf 1}_n \rangle^{-1} \exp(Ax) - b \|_2^2$ and Resscaled version $\| \exp(Ax) - \langle \exp(Ax), {\bf 1}_n \rangle b \|_2^2 $ In contrast to previous approaches, we adopt a vectorization technique to address the regression problem in matrix formulation. This approach expands the dimension from $d$ to $d^2$, resembling the formulation of the regression problem mentioned earlier. Upon completing the lipschitz analysis of our regression function, we have derived our main result concerning in-context learning.

Pre-trained language models (PLMs) have shown unprecedented potential in various fields, especially as the backbones for question-answering (QA) systems. However, they tend to be easily deceived by tricky questions such as "How many eyes does the sun have?". Such frailties of PLMs often allude to the lack of knowledge within them. In this paper, we find that the PLMs already possess the knowledge required to rebut such questions, and the key is how to activate the knowledge. To systematize this observation, we investigate the PLMs' responses to one kind of tricky questions, i.e., the false premises questions (FPQs). We annotate a FalseQA dataset containing 2365 human-written FPQs, with the corresponding explanations for the false premises and the revised true premise questions. Using FalseQA, we discover that PLMs are capable of discriminating FPQs by fine-tuning on moderate numbers (e.g., 256) of examples. PLMs also generate reasonable explanations for the false premise, which serve as rebuttals. Further replaying a few general questions during training allows PLMs to excel on FPQs and general questions simultaneously. Our work suggests that once the rebuttal ability is stimulated, knowledge inside the PLMs can be effectively utilized to handle FPQs, which incentivizes the research on PLM-based QA systems.

Understanding the causal relationships that underlie a system is a fundamental prerequisite to accurate decision-making. In this work, we explore how expert knowledge can be used to improve the data-driven identification of causal graphs, beyond Markov equivalence classes. In doing so, we consider a setting where we can query an expert about the orientation of causal relationships between variables, but where the expert may provide erroneous information. We propose strategies for amending such expert knowledge based on consistency properties, e.g., acyclicity and conditional independencies in the equivalence class. We then report a case study, on real data, where a large language model is used as an imperfect expert.

We are amidst an explosion of artificial intelligence research, particularly around large language models (LLMs). These models have a range of applications across domains like medicine, finance, commonsense knowledge graphs, and crowdsourcing. Investigation into LLMs as part of crowdsourcing workflows remains an under-explored space. The crowdsourcing research community has produced a body of work investigating workflows and methods for managing complex tasks using hybrid human-AI methods. Within crowdsourcing, the role of LLMs can be envisioned as akin to a cog in a larger wheel of workflows. From an empirical standpoint, little is currently understood about how LLMs can improve the effectiveness of crowdsourcing workflows and how such workflows can be evaluated. In this work, we present a vision for exploring this gap from the perspectives of various stakeholders involved in the crowdsourcing paradigm -- the task requesters, crowd workers, platforms, and end-users. We identify junctures in typical crowdsourcing workflows at which the introduction of LLMs can play a beneficial role and propose means to augment existing design patterns for crowd work.

For instance, studies demonstrate that ChatGPT exceeds the performance of crowd-workers in tasks encompassing relevance, stance, sentiment, topic identification, and frame detection (Gilardi, Alizadeh and Kubli, 2023), that it outperforms trained annotators in detecting the political party affiliations of Twitter users (Törnberg, 2023), and that it achieves accuracy scores over 0.6 for tasks such as stance, sentiment, hate speech detection, and bot identification (Zhu et al., 2023). Notably, ChatGPT also demonstrates the ability to correctly classify more than 70% of news as either true or false (Hoes, Altay and Bermeo, 2023), which suggests that LLMs might potentially be used to assist content moderation processes. While the performance of LLMs for text annotation is promising, there are several aspects that remain unclear and require further research. Among these is the impact of different approaches such as zero-shot versus few-shot learning and settings such as varying temperature parameters. Zero-shot learning allows models to predict for unseen tasks, while few-shot learning uses a small number of examples to generalize to new tasks. The conditions under which one approach outperforms the other are not fully understood yet.

The rapid development of online recruitment services has encouraged the utilization of recommender systems to streamline the job seeking process. Predominantly, current job recommendations deploy either collaborative filtering or person-job matching strategies. However, these models tend to operate as "black-box" systems and lack the capacity to offer explainable guidance to job seekers. Moreover, conventional matching-based recommendation methods are limited to retrieving and ranking existing jobs in the database, restricting their potential as comprehensive career AI advisors. To this end, here we present GIRL (GeneratIve job Recommendation based on Large language models), a novel approach inspired by recent advancements in the field of Large Language Models (LLMs). We initially employ a Supervised Fine-Tuning (SFT) strategy to instruct the LLM-based generator in crafting suitable Job Descriptions (JDs) based on the Curriculum Vitae (CV) of a job seeker. Moreover, we propose to train a model which can evaluate the matching degree between CVs and JDs as a reward model, and we use Proximal Policy Optimization (PPO)-based Reinforcement Learning (RL) method to further fine-tine the generator. This aligns the generator with recruiter feedback, tailoring the output to better meet employer preferences. In particular, GIRL serves as a job seeker-centric generative model, providing job suggestions without the need of a candidate set. This capability also enhances the performance of existing job recommendation models by supplementing job seeking features with generated content. With extensive experiments on a large-scale real-world dataset, we demonstrate the substantial effectiveness of our approach. We believe that GIRL introduces a paradigm-shifting approach to job recommendation systems, fostering a more personalized and comprehensive job-seeking experience.

In recent years, the barriers to the development of Artificial Intelligence (AI) have been broken down with the rapid progress of ABC technologies in computing: AI, Big Data, and Cloud Computing, as well as the emergence of cost-effective specialized hardware [213] and software [98]. This has led to the world entering the third wave of AI development: Deep Learning [117]. The success of current data-driven AI relies on massive amounts of training data and follows a gather-and-analyze paradigm [233], which confronts with challenges of complying with rigorous data protection regulations such as OECD Privacy Guidelines [217] and General and Data Protection Regulation (GDPR) [223]. So although data-centric AI is currently mainstream paradigm, Federated Learning [132], a novel model-centric distributed collaborative training framework, is gaining popularity in both academia and industry for its advantages in complying with privacy regulations [219]. According to the definitions of IEEE Standard for Federated Machine Learning (FML, aka FL) [205], FL is a framework or system that enables multiple participants to collaboratively build and use machine learning models without disclosing the raw and private data owned by the participants while achieving good performance. For example, a typical workflow of FL systems is that the entity with modeling demand (aka FL server) first deploys the FL services and initializes the model training task, and then distributing this task to participants with training data (aka FL clients) for modeling [13]. Based on this workflow pattern, many FL frameworks have been derived with specialized improvements in communication [111, 161, 240], optimizaiton [107, 129, 133], robustness [44, 124, 198] and privacy [14, 32, 62]. While these fascinating improvements greatly enhance the utility of FL, they all follow a task-based interaction paradigm, in which an FL server dominates the cooperation between FL participants. In this narrow interpretation of FL, the data owner is treated more like a worker than a collaborator and performs training primarily for the benefit of the server's goals.

Text is by far the most ubiquitous source of knowledge and information and should be made easily accessible to as many people as possible; however, texts often contain complex words that hinder reading comprehension and accessibility. Therefore, suggesting simpler alternatives for complex words without compromising meaning would help convey the information to a broader audience. This paper proposes mTLS, a multilingual controllable Transformer-based Lexical Simplification (LS) system fined-tuned with the T5 model. The novelty of this work lies in the use of language-specific prefixes, control tokens, and candidates extracted from pre-trained masked language models to learn simpler alternatives for complex words. The evaluation results on three well-known LS datasets -- LexMTurk, BenchLS, and NNSEval -- show that our model outperforms the previous state-of-the-art models like LSBert and ConLS. Moreover, further evaluation of our approach on the part of the recent TSAR-2022 multilingual LS shared-task dataset shows that our model performs competitively when compared with the participating systems for English LS and even outperforms the GPT-3 model on several metrics. Moreover, our model obtains performance gains also for Spanish and Portuguese.