Important tasks such as reasoning and planning are fundamentally algorithmic, meaning that solving them robustly requires acquiring true reasoning or planning algorithms, rather than shortcuts. Large Language Models lack true algorithmic ability primarily because of the limitations of neural network optimization algorithms, their optimization data and optimization objective, but also due to architectural inexpressivity. To solve this, our paper proposes augmenting LLMs with a library of fundamental operations and sophisticated differentiable programs, so that common algorithms do not need to be learned from scratch. We add memory, registers, basic operations, and adaptive recurrence to a transformer architecture built on LLaMA3. Then, we define a method for directly compiling algorithms into a differentiable starting library, which is used natively and propagates gradients for optimization. In this preliminary study, we explore the feasability of augmenting LLaMA3 with a differentiable computer, for instance by fine-tuning small transformers on simple algorithmic tasks with variable computational depth.

Language Models (LMs) have shown their application for tasks pertinent to code and several code~LMs have been proposed recently. The majority of the studies in this direction only focus on the improvements in performance of the LMs on different benchmarks, whereas LMs are considered black boxes. Besides this, a handful of works attempt to understand the role of attention layers in the code~LMs. Nonetheless, feed-forward layers remain under-explored which consist of two-thirds of a typical transformer model's parameters. In this work, we attempt to gain insights into the inner workings of code language models by examining the feed-forward layers. To conduct our investigations, we use two state-of-the-art code~LMs, Codegen-Mono and Ploycoder, and three widely used programming languages, Java, Go, and Python. We focus on examining the organization of stored concepts, the editability of these concepts, and the roles of different layers and input context size variations for output generation. Our empirical findings demonstrate that lower layers capture syntactic patterns while higher layers encode abstract concepts and semantics. We show concepts of interest can be edited within feed-forward layers without compromising code~LM performance. Additionally, we observe initial layers serve as ``thinking'' layers, while later layers are crucial for predicting subsequent code tokens. Furthermore, we discover earlier layers can accurately predict smaller contexts, but larger contexts need critical later layers' contributions. We anticipate these findings will facilitate better understanding, debugging, and testing of code~LMs.

Large Language Models (LLMs) excel in various applications, including text generation and complex tasks. However, the misuse of LLMs raises concerns about the authenticity and ethical implications of the content they produce, such as deepfake news, academic fraud, and copyright infringement. Watermarking techniques, which embed identifiable markers in machine-generated text, offer a promising solution to these issues by allowing for content verification and origin tracing. Unfortunately, the robustness of current LLM watermarking schemes under potential watermark removal attacks has not been comprehensively explored. In this paper, to fill this gap, we first systematically comb the mainstream watermarking schemes and removal attacks on machine-generated texts, and then we categorize them into pre-text (before text generation) and post-text (after text generation) classes so that we can conduct diversified analyses. In our experiments, we evaluate eight watermarks (five pre-text, three post-text) and twelve attacks (two pre-text, ten post-text) across 87 scenarios. Evaluation results indicate that (1) KGW and Exponential watermarks offer high text quality and watermark retention but remain vulnerable to most attacks; (2) Post-text attacks are found to be more efficient and practical than pre-text attacks; (3) Pre-text watermarks are generally more imperceptible, as they do not alter text fluency, unlike post-text watermarks; (4) Additionally, combined attack methods can significantly increase effectiveness, highlighting the need for more robust watermarking solutions. Our study underscores the vulnerabilities of current techniques and the necessity for developing more resilient schemes.

We propose Waymo Open Motion Dataset-Reasoning (WOMD-Reasoning), a language annotation dataset built on WOMD, with a focus on describing and reasoning interactions and intentions in driving scenarios. Previous language datasets primarily captured interactions caused by close distances. However, interactions induced by traffic rules and human intentions, which can occur over long distances, are yet sufficiently covered, despite being very common and more challenging for prediction or planning models to understand. Therefore, our WOMD-Reasoning focuses extensively on these interactions, providing a total of 409k Q&As for varying types of interactions. Additionally, WOMD-Reasoning presents by far the largest Q&A dataset on real-world driving scenarios, with around 3 million Q&As covering various topics of autonomous driving from map descriptions, motion status descriptions, to narratives and analyses of agents' interactions, behaviors, and intentions. This extensive textual information enables fine-tuning driving-related Large Language Models (LLMs) for a wide range of applications like scene description, prediction, planning, etc. By incorporating interaction and intention language from WOMD-Reasoning, we see significant enhancements in the performance of the state-of-the-art trajectory prediction model, Multipath++, with improvements of 10.14% in $MR_6$ and 6.90% in $minFDE_6$, proving the effectiveness of WOMD-Reasoning. We hope WOMD-Reasoning would empower LLMs in driving to offer better interaction understanding and behavioral reasoning. The dataset is available on https://waymo.com/open/download .

Large Language Models (LLMs) have performed exceptionally in various text-generative tasks, including question answering, translation, code completion, etc. However, the over-assistance of LLMs has raised the challenge of "jailbreaking", which induces the model to generate malicious responses against the usage policy and society by designing adversarial prompts. With the emergence of jailbreak attack methods exploiting different vulnerabilities in LLMs, the corresponding safety alignment measures are also evolving. In this paper, we propose a comprehensive and detailed taxonomy of jailbreak attack and defense methods. For instance, the attack methods are divided into black-box and white-box attacks based on the transparency of the target model. Meanwhile, we classify defense methods into prompt-level and model-level defenses. Additionally, we further subdivide these attack and defense methods into distinct sub-classes and present a coherent diagram illustrating their relationships. We also conduct an investigation into the current evaluation methods and compare them from different perspectives. Our findings aim to inspire future research and practical implementations in safeguarding LLMs against adversarial attacks. Above all, although jailbreak remains a significant concern within the community, we believe that our work enhances the understanding of this domain and provides a foundation for developing more secure LLMs.

We present a new method for large language models to solve compositional tasks. Although they have shown strong performance on traditional language understanding tasks, large language models struggle to solve compositional tasks, where the solution depends on solving smaller instances of the same problem. We propose a natural approach to solve compositional tasks recursively. Our method, Re-Tuning, tunes models to break down a problem into subproblems, solve those subproblems, and combine the results. We show that our method significantly improves model performance on three representative compositional tasks: integer addition, dynamic programming, and parity. Compared to state-of-the-art methods that keep intermediate steps towards solving the problems, Re-Tuning achieves significantly higher accuracy and is more GPU memory efficient.

We introduce PoPreRo, the first dataset for Popularity Prediction of Romanian posts collected from Reddit. The PoPreRo dataset includes a varied compilation of post samples from five distinct subreddits of Romania, totaling 28,107 data samples. Along with our novel dataset, we introduce a set of competitive models to be used as baselines for future research. Interestingly, the top-scoring model achieves an accuracy of 61.35% and a macro F1 score of 60.60% on the test set, indicating that the popularity prediction task on PoPreRo is very challenging. Further investigations based on few-shot prompting the Falcon-7B Large Language Model also point in the same direction. We thus believe that PoPreRo is a valuable resource that can be used to evaluate models on predicting the popularity of social media posts in Romanian. We release our dataset at https://github.com/ana-rogoz/PoPreRo.

Massive Open Online Courses (MOOCs) have significantly enhanced educational accessibility by offering a wide variety of courses and breaking down traditional barriers related to geography, finance, and time. However, students often face difficulties navigating the vast selection of courses, especially when exploring new fields of study. Driven by this challenge, researchers have been exploring course recommender systems to offer tailored guidance that aligns with individual learning preferences and career aspirations. These systems face particular challenges in effectively addressing the ``cold start'' problem for new users. Recent advancements in recommender systems suggest integrating large language models (LLMs) into the recommendation process to enhance personalized recommendations and address the ``cold start'' problem. Motivated by these advancements, our study introduces RAMO (Retrieval-Augmented Generation for MOOCs), a system specifically designed to overcome the ``cold start'' challenges of traditional course recommender systems. The RAMO system leverages the capabilities of LLMs, along with Retrieval-Augmented Generation (RAG)-facilitated contextual understanding, to provide course recommendations through a conversational interface, aiming to enhance the e-learning experience.

We introduce the sequence classification problem CIViC Evidence to the field of medical NLP. CIViC Evidence denotes the multi-label classification problem of assigning labels of clinical evidence to abstracts of scientific papers which have examined various combinations of genomic variants, cancer types, and treatment approaches. We approach CIViC Evidence using different language models: We fine-tune pretrained checkpoints of BERT and RoBERTa on the CIViC Evidence dataset and challenge their performance with models of the same architecture which have been pretrained on domain-specific text. In this context, we find that BiomedBERT and BioLinkBERT can outperform BERT on CIViC Evidence (+0.8% and +0.9% absolute improvement in class-support weighted F1 score). All transformer-based models show a clear performance edge when compared to a logistic regression trained on bigram tf-idf scores (+1.5 - 2.7% improved F1 score). We compare the aforementioned BERT-like models to OpenAI's GPT-4 in a few-shot setting (on a small subset of our original test dataset), demonstrating that, without additional prompt-engineering or fine-tuning, GPT-4 performs worse on CIViC Evidence than our six fine-tuned models (66.1% weighted F1 score compared to 71.8% for the best fine-tuned model). However, performance gets reasonably close to the benchmark of a logistic regression model trained on bigram tf-idf scores (67.7% weighted F1 score).

The recent advancements in large language models (LLMs) with billions of parameters have significantly boosted their performance across various real-world applications. However, the inference processes for these models require substantial energy and computational resources, presenting considerable deployment challenges. In contrast, human brains, which contain approximately 86 billion biological neurons, exhibit significantly greater energy efficiency compared to LLMs with a similar number of parameters. Inspired by this, we redesign 7 to 70 billion parameter LLMs using bio-plausible spiking mechanisms, emulating the efficient behavior of the human brain. We propose the first spiking large language model as recent LLMs termed SpikeLLM. Coupled with the proposed model, a novel spike-driven quantization framework named Optimal Brain Spiking is introduced to reduce the energy cost and accelerate inference speed via two essential approaches: first (second)-order differentiation-based salient channel detection, and per-channel salient outlier expansion with Generalized Integrate-and-Fire neurons. Our proposed spike-driven quantization can plug in main streams of quantization training methods. In the OmniQuant pipeline, SpikeLLM significantly reduces 25.51% WikiText2 perplexity and improves 3.08% average accuracy of 6 zero-shot datasets on a LLAMA2-7B 4A4W model. In the GPTQ pipeline, SpikeLLM realizes a sparse ternary quantization, which achieves additive in all linear layers. Compared with PB-LLM with similar operations, SpikeLLM also exceeds significantly. We will release our code on GitHub.