Large language models exhibit surprising emergent generalization properties, yet also struggle on many simple reasoning tasks such as arithmetic and parity. This raises the question of if and when Transformer models can learn the true algorithm for solving a task. We study the scope of Transformers' abilities in the specific setting of length generalization on algorithmic tasks. Here, we propose a unifying framework to understand when and how Transformers can exhibit strong length generalization on a given task. Specifically, we leverage RASP (Weiss et al., 2021) -- a programming language designed for the computational model of a Transformer -- and introduce the RASP-Generalization Conjecture: Transformers tend to length generalize on a task if the task can be solved by a short RASP program which works for all input lengths. This simple conjecture remarkably captures most known instances of length generalization on algorithmic tasks. Moreover, we leverage our insights to drastically improve generalization performance on traditionally hard tasks (such as parity and addition). On the theoretical side, we give a simple example where the "min-degree-interpolator" model of learning from Abbe et al. (2023) does not correctly predict Transformers' out-of-distribution behavior, but our conjecture does. Overall, our work provides a novel perspective on the mechanisms of compositional generalization and the algorithmic capabilities of Transformers.

Safeguards designed to prevent OpenAI's GPT-4 artificial intelligence from answering harmful prompts failed when it received requests in languages such as Scots Gaelic or Zulu. This allowed researchers to get AI-generated answers on how to build a homemade bomb or perform insider trading. The vulnerability demonstrated in the large language model involves instructing the AI in languages that are mostly absent from its training data.

If you've got a story to tell or advice to give, an eBook is a compelling avenue to share with the world and maybe even make a little money. You don't even have to work all that hard thanks to My AI eBook Creation Pro. Powered by ChatGPT, this eBook creation tool takes all of the pain points out of the process of creating an eBook, and it's on sale for just $34.99 now. With My AI eBook, you can go from concept to published eBook faster than ever. You don't need any technical background thanks to the user-friendly interface that guides you through putting out your ideas and polishing your content.

The UK government, with its reversals on climate policy and commitment to oil drilling and air pollution, usually seems to be pro-apocalypse. But lately, senior British politicians have been on a save-the-world tour. Prime minister Rishi Sunak, his ministers, and diplomats have been briefing their international counterparts about the existential dangers of runaway artificial superintelligence, which, they warn, could engineer bioweapons, empower autocrats, undermine democracy, and threaten the financial system. "I do not believe we can hold back the tide," deputy prime minister Oliver Dowden told the United Nations in late September. Dowden's doomerism is supposed to drum up support for the UK government's global summit on AI governance, scheduled for November 1 and 2. The event is being billed as the moment that the tide turns on the specter of killer AI, a chance to start building international consensus toward mitigating that risk.

This paper introduces a new IncidentAI dataset for safety prevention. Different from prior corpora that usually contain a single task, our dataset comprises three tasks: named entity recognition, cause-effect extraction, and information retrieval. The dataset is annotated by domain experts who have at least six years of practical experience as high-pressure gas conservation managers. We validate the contribution of the dataset in the scenario of safety prevention. Preliminary results on the three tasks show that NLP techniques are beneficial for analyzing incident reports to prevent future failures. The dataset facilitates future research in NLP and incident management communities. The access to the dataset is also provided (the IncidentAI dataset is available at: https://github.com/Cinnamon/incident-ai-dataset).

Recent progress in natural language processing (NLP) owes much to remarkable advances in large language models (LLMs). Nevertheless, LLMs frequently "hallucinate," resulting in non-factual outputs. Our carefully designed human evaluation substantiates the serious hallucination issue, revealing that even GPT-3.5 produces factual outputs less than 25% of the time. This underscores the importance of fact verifiers in order to measure and incentivize progress. Our systematic investigation affirms that LLMs can be repurposed as effective fact verifiers with strong correlations with human judgments, at least in the Wikipedia domain. Surprisingly, FLAN-T5-11B, the least factual generator in our study, performs the best as a fact verifier, even outperforming more capable LLMs like GPT3.5 and ChatGPT. Delving deeper, we analyze the reliance of these LLMs on high-quality evidence, as well as their deficiencies in robustness and generalization ability. Our study presents insights for developing trustworthy generation models.

In this work, we investigate the capacity of language models to generate explicit, interpretable, and interactive world models of scientific and common-sense reasoning tasks. We operationalize this as a task of generating text games, expressed as hundreds of lines of Python code. To facilitate this task, we introduce ByteSized32 (Code: github.com/cognitiveailab/BYTESIZED32), a corpus of 32 reasoning-focused text games totaling 20k lines of Python code. We empirically demonstrate that GPT-4 can use these games as templates for single-shot in-context learning, successfully producing runnable games on unseen topics in 28% of cases. When allowed to self-reflect on program errors, game runnability substantially increases to 57%. While evaluating simulation fidelity is labor-intensive, we introduce a suite of automated metrics to assess game fidelity, technical validity, adherence to task specifications, and winnability, showing a high degree of agreement with expert human ratings. We pose this as a challenge task to spur further development at the juncture of world modeling and code generation.

Current approaches in paraphrase generation and detection heavily rely on a single general similarity score, ignoring the intricate linguistic properties of language. This paper introduces two new tasks to address this shortcoming by considering paraphrase types - specific linguistic perturbations at particular text positions. We name these tasks Paraphrase Type Generation and Paraphrase Type Detection. Our results suggest that while current techniques perform well in a binary classification scenario, i.e., paraphrased or not, the inclusion of fine-grained paraphrase types poses a significant challenge. While most approaches are good at generating and detecting general semantic similar content, they fail to understand the intrinsic linguistic variables they manipulate. Models trained in generating and identifying paraphrase types also show improvements in tasks without them. In addition, scaling these models further improves their ability to understand paraphrase types. We believe paraphrase types can unlock a new paradigm for developing paraphrase models and solving tasks in the future.

Despite the impressive performance achieved by pre-trained language-and-vision models in downstream tasks, it remains an open question whether this reflects a proper understanding of image-text interaction. In this work, we explore to what extent they handle basic linguistic constructions -- active-passive voice, coordination, and relative clauses -- that even preschool children can typically master. We present BLA, a novel, automatically constructed benchmark to evaluate multimodal models on these Basic Language Abilities. We show that different types of Transformer-based systems, such as CLIP, ViLBERT, and BLIP2, generally struggle with BLA in a zero-shot setting, in line with previous findings. Our experiments, in particular, show that most of the tested models only marginally benefit when fine-tuned or prompted with construction-specific samples. Yet, the generative BLIP2 shows promising trends, especially in an in-context learning setting. This opens the door to using BLA not only as an evaluation benchmark but also to improve models' basic language abilities.

We study continual event extraction, which aims to extract incessantly emerging event information while avoiding forgetting. We observe that the semantic confusion on event types stems from the annotations of the same text being updated over time. The imbalance between event types even aggravates this issue. This paper proposes a novel continual event extraction model with semantic confusion rectification. We mark pseudo labels for each sentence to alleviate semantic confusion. We transfer pivotal knowledge between current and previous models to enhance the understanding of event types. Moreover, we encourage the model to focus on the semantics of long-tailed event types by leveraging other associated types. Experimental results show that our model outperforms state-of-the-art baselines and is proficient in imbalanced datasets.