Large Language Models (LLMs) with billions of parameters have drastically transformed AI applications. However, their demanding computation during inference has raised significant challenges for deployment on resource-constrained devices. Despite recent trends favoring alternative activation functions such as GELU or SiLU, known for increased computation, this study strongly advocates for reinstating ReLU activation in LLMs. We demonstrate that using the ReLU activation function has a negligible impact on convergence and performance while significantly reducing computation and weight transfer. This reduction is particularly valuable during the memory-bound inference step, where efficiency is paramount. Exploring sparsity patterns in ReLU-based LLMs, we unveil the reutilization of activated neurons for generating new tokens and leveraging these insights, we propose practical strategies to substantially reduce LLM inference computation up to three times, using ReLU activations with minimal performance trade-offs.

Graphs are a powerful tool for representing and analyzing complex relationships in real-world applications such as social networks, recommender systems, and computational finance. Reasoning on graphs is essential for drawing inferences about the relationships between entities in a complex system, and to identify hidden patterns and trends. Despite the remarkable progress in automated reasoning with natural text, reasoning on graphs with large language models (LLMs) remains an understudied problem. In this work, we perform the first comprehensive study of encoding graph-structured data as text for consumption by LLMs. We show that LLM performance on graph reasoning tasks varies on three fundamental levels: (1) the graph encoding method, (2) the nature of the graph task itself, and (3) interestingly, the very structure of the graph considered. These novel results provide valuable insight on strategies for encoding graphs as text. Using these insights we illustrate how the correct choice of encoders can boost performance on graph reasoning tasks inside LLMs by 4.8% to 61.8%, depending on the task.

Test stimuli generation has been a crucial but labor-intensive task in hardware design verification. In this paper, we revolutionize this process by harnessing the power of large language models (LLMs) and present a novel benchmarking framework, LLM4DV. This framework introduces a prompt template for interactively eliciting test stimuli from the LLM, along with four innovative prompting improvements to support the pipeline execution and further enhance its performance. We compare LLM4DV to traditional constrained-random testing (CRT), using three self-designed design-under-test (DUT) modules. Experiments demonstrate that LLM4DV excels in efficiently handling straightforward DUT scenarios, leveraging its ability to employ basic mathematical reasoning and pre-trained knowledge. While it exhibits reduced efficiency in complex task settings, it still outperforms CRT in relative terms. The proposed framework and the DUT modules used in our experiments will be open-sourced upon publication.

Embeddings have become a pivotal means to represent complex, multi-faceted information about entities, concepts, and relationships in a condensed and useful format. Nevertheless, they often preclude direct interpretation. While downstream tasks make use of these compressed representations, meaningful interpretation usually requires visualization using dimensionality reduction or specialized machine learning interpretability methods. This paper addresses the challenge of making such embeddings more interpretable and broadly useful, by employing Large Language Models (LLMs) to directly interact with embeddings -- transforming abstract vectors into understandable narratives. By injecting embeddings into LLMs, we enable querying and exploration of complex embedding data. We demonstrate our approach on a variety of diverse tasks, including: enhancing concept activation vectors (CAVs), communicating novel embedded entities, and decoding user preferences in recommender systems. Our work couples the immense information potential of embeddings with the interpretative power of LLMs.

Preventing the performance decay of Transformers on inputs longer than those used for training has been an important challenge in extending the context length of these models. Though the Transformer architecture has fundamentally no limits on the input sequence lengths it can process, the choice of position encoding used during training can limit the performance of these models on longer inputs. We propose a novel functional relative position encoding with progressive interpolation, FIRE, to improve Transformer generalization to longer contexts. We theoretically prove that this can represent some of the popular relative position encodings, such as T5's RPE, Alibi, and Kerple. We next empirically show that FIRE models have better generalization to longer contexts on both zero-shot language modeling and long text benchmarks.

Weight decay is a broadly used technique for training state-of-the-art deep networks, including large language models. Despite its widespread usage, its role remains poorly understood. In this work, we highlight that the role of weight decay in modern deep learning is different from its regularization effect studied in classical learning theory. For overparameterized deep networks, we show how weight decay modifies the optimization dynamics enhancing the ever-present implicit regularization of SGD via the loss stabilization mechanism. In contrast, for underparameterized large language models trained with nearly online SGD, we describe how weight decay balances the bias-variance tradeoff in stochastic optimization leading to lower training loss. Moreover, we show that weight decay also prevents sudden loss divergences for bfloat16 mixed-precision training which is a crucial tool for LLM training. Overall, we present a unifying perspective from ResNets on vision tasks to LLMs: weight decay is never useful as an explicit regularizer but instead changes the training dynamics in a desirable way. Weight decay serves to constrain the network capacity (Goodfellow et al., 2016) and acts as a mechanism for suppressing irrelevant weight components, aligning with the principles of Occam's razor (Krogh & Hertz, 1991). It is central in discussions on generalization bounds (Shalev-Shwartz & Ben-David, 2014), albeit a recent empirical study by Jiang et al. (2020) casts doubt on how well norm-based measures correlate with generalization for deep networks. Weight decay is also known to yield a regularization of the input-output Jacobian (Zhang et al., 2018) and to alter the training dynamics of scale-invariant networks by changing the effective learning rate (Van Laarhoven, 2017). Weight decay is widely used for training most state-of-theart deep networks such as GPT-3 (Brown et al., 2020), CLIP (Radford et al., 2021), or PALM (Chowdhery et al., 2022). We argue that despite its widespread usage, its effect is still poorly understood: in some cases it acts as a regularizer but in some cases as a tool for better optimization. Although the regularization effect of weight decay is thoroughly studied in classical learning theory, deep networks are already equipped with strong implicit regularization coming from the parameter initialization, optimization algorithm, and architecture (Zhang et al., 2016). Moreover, recent years have brought along new architectures and settings such as transformers (Vaswani et al., 2017) and nearly one-epoch language modelling (Brown et al., 2020; Hoffmann et al., 2022).

Text retrieval plays a crucial role in incorporating factual knowledge for decision making into language processing pipelines, ranging from chat-based web search to question answering systems. Current state-of-the-art text retrieval models leverage pre-trained large language models (LLMs) to achieve competitive performance, but training LLM-based retrievers via typical contrastive losses requires intricate heuristics, including selecting hard negatives and using additional supervision as learning signals. This reliance on heuristics stems from the fact that the contrastive loss itself is heuristic and does not directly optimize the downstream metrics of decision quality at the end of the processing pipeline. To address this issue, we introduce Neural PG-RANK, a novel training algorithm that learns to rank by instantiating a LLM as a Plackett-Luce ranking policy. Neural PG-RANK provides a principled method for end-to-end training of retrieval models as part of larger decision systems via policy gradient, with little reliance on complex heuristics, and it effectively unifies the training objective with downstream decision-making quality. We conduct extensive experiments on various text retrieval benchmarks. The results demonstrate that when the training objective aligns with the evaluation setup, Neural PG-RANK yields remarkable in-domain performance improvement, with substantial out-of-domain generalization to some critical datasets employed in downstream question answering tasks.

Autoregressive large language models (LLMs) compress knowledge from their training data through next-token conditional distributions. This limits tractable querying of this knowledge to start-to-end autoregressive sampling. However, many tasks of interest -- including sequence continuation, infilling, and other forms of constrained generation -- involve sampling from intractable posterior distributions. We address this limitation by using amortized Bayesian inference to sample from these intractable posteriors. Such amortization is algorithmically achieved by fine-tuning LLMs via diversity-seeking reinforcement learning algorithms: generative flow networks (GFlowNets). We empirically demonstrate that this distribution-matching paradigm of LLM fine-tuning can serve as an effective alternative to maximum-likelihood training and reward-maximizing policy optimization. As an important application, we interpret chain-of-thought reasoning as a latent variable modeling problem and demonstrate that our approach enables data-efficient adaptation of LLMs to tasks that require multi-step rationalization and tool use.

In what sense does a large language model have knowledge? The answer to this question extends beyond the capabilities of a particular AI system, and challenges our assumptions about the nature of knowledge and intelligence. We answer by granting LLMs "instrumental knowledge"; knowledge defined by a certain set of abilities. We then ask how such knowledge is related to the more ordinary, "worldly" knowledge exhibited by human agents, and explore this in terms of the degree to which instrumental knowledge can be said to incorporate the structured world models of cognitive science. We discuss ways LLMs could recover degrees of worldly knowledge, and suggest such recovery will be governed by an implicit, resource-rational tradeoff between world models and task demands.

Recent advancements in Large Language Models (LLMs) have demonstrated impressive capabilities across a range of natural language processing tasks, especially in reasoning, a cornerstone for achieving Artificial General Intelligence (AGI). However, commonly used benchmarks may not fully encapsulate the inferential abilities of these models in real-world scenarios. To address this gap, a new form of Question-Answering (QA) task, termed Reasoning with Redundant Information Provided (RRIP), is introduced. The study designed a modified version of the grade school math 8K (GSM-8K) dataset which has several variants focusing on different attributes of redundant information. This investigation evaluates two popular LLMs, LlaMA2-13B-chat and generative pre-trained transformer 3.5 (GPT-3.5), contrasting their performance on traditional QA tasks against the RRIP tasks. Findings indicate that while these models achieved moderate success on standard QA benchmarks, their performance notably declines when assessed on RRIP tasks. The study not only highlights the limitations of current LLMs in handling redundant information but also suggests that future training of these models should focus on incorporating redundant information into the training data to increase the performance on RRIP tasks.