In this paper, we investigate the output token probability information in the output embedding of language models. We provide an approximate common log-linear encoding of output token probabilities within the output embedding vectors and demonstrate that it is accurate and sparse when the output space is large and output logits are concentrated. Based on such findings, we edit the encoding in output embedding to modify the output probability distribution accurately. Moreover, the sparsity we find in output probability encoding suggests that a large number of dimensions in the output embedding do not contribute to causal language modeling. Therefore, we attempt to delete the output-unrelated dimensions and find more than 30% of the dimensions can be deleted without significant movement in output distribution and degeneration on sequence generation. Additionally, in training dynamics, we use such encoding as a probe and find that the output embeddings capture token frequency information in early steps, even before an obvious convergence starts.

Large language models (LLMs) have shown remarkable performance across a wide range of applications, often outperforming human experts. However, deploying these parameter-heavy models efficiently for diverse inference use cases requires carefully designed hardware platforms with ample computing, memory, and network resources. With LLM deployment scenarios and models evolving at breakneck speed, the hardware requirements to meet SLOs remains an open research question. In this work, we present an analytical tool, GenZ, to study the relationship between LLM inference performance and various platform design parameters. Our analysis provides insights into configuring platforms for different LLM workloads and use cases. We quantify the platform requirements to support SOTA LLMs models like LLaMA and GPT-4 under diverse serving settings. Furthermore, we project the hardware capabilities needed to enable future LLMs potentially exceeding hundreds of trillions of parameters. The trends and insights derived from GenZ can guide AI engineers deploying LLMs as well as computer architects designing next-generation hardware accelerators and platforms. Ultimately, this work sheds light on the platform design considerations for unlocking the full potential of large language models across a spectrum of applications. The source code is available at https://github.com/abhibambhaniya/GenZ-LLM-Analyzer .

This paper introduces Helix, a distributed system for high-throughput, low-latency large language model (LLM) serving on heterogeneous GPU clusters. A key idea behind Helix is to formulate inference computation of LLMs over heterogeneous GPUs and network connections as a max-flow problem for a directed, weighted graph, whose nodes represent GPU instances and edges capture both GPU and network heterogeneity through their capacities. Helix then uses a mixed integer linear programming (MILP) algorithm to discover highly optimized strategies to serve LLMs. This approach allows Helix to jointly optimize model placement and request scheduling, two highly entangled tasks in heterogeneous LLM serving. Our evaluation on several heterogeneous cluster settings ranging from 24 to 42 GPU nodes shows that Helix improves serving throughput by up to 2.7$\times$ and reduces prompting and decoding latency by up to 2.8$\times$ and 1.3$\times$, respectively, compared to best existing approaches.

Test generation has been a critical and labor-intensive process in hardware design verification. Recently, the emergence of Large Language Model (LLM) with their advanced understanding and inference capabilities, has introduced a novel approach. In this work, we investigate the integration of LLM into the Coverage Directed Test Generation (CDG) process, where the LLM functions as a Verilog Reader. It accurately grasps the code logic, thereby generating stimuli that can reach unexplored code branches. We compare our framework with random testing, using our self-designed Verilog benchmark suite. Experiments demonstrate that our framework outperforms random testing on designs within the LLM's comprehension scope. Our work also proposes prompt engineering optimizations to augment LLM's understanding scope and accuracy.

Conversational grounding is a collaborative mechanism for establishing mutual knowledge among participants engaged in a dialogue. This experimental study analyzes information-seeking conversations to investigate the capabilities of large language models in classifying dialogue turns related to explicit or implicit grounding and predicting grounded knowledge elements. Our experimental results reveal challenges encountered by large language models in the two tasks and discuss ongoing research efforts to enhance large language model-based conversational grounding comprehension through pipeline architectures and knowledge bases. These initiatives aim to develop more effective dialogue systems that are better equipped to handle the intricacies of grounded knowledge in conversations.

We explore how interaction with large language models (LLMs) can give rise to emergent behaviors, empowering players to participate in the evolution of game narratives. Our testbed is a text-adventure game in which players attempt to solve a mystery under a fixed narrative premise, but can freely interact with non-player characters generated by GPT-4, a large language model. We recruit 28 gamers to play the game and use GPT-4 to automatically convert the game logs into a node-graph representing the narrative in the player's gameplay. We find that through their interactions with the non-deterministic behavior of the LLM, players are able to discover interesting new emergent nodes that were not a part of the original narrative but have potential for being fun and engaging. Players that created the most emergent nodes tended to be those that often enjoy games that facilitate discovery, exploration and experimentation.

We consider the task of active geo-localization (AGL) in which an agent uses a sequence of visual cues observed during aerial navigation to find a target specified through multiple possible modalities. This could emulate a UAV involved in a search-and-rescue operation navigating through an area, observing a stream of aerial images as it goes. The AGL task is associated with two important challenges. Firstly, an agent must deal with a goal specification in one of multiple modalities (e.g., through a natural language description) while the search cues are provided in other modalities (aerial imagery). The second challenge is limited localization time (e.g., limited battery life, urgency) so that the goal must be localized as efficiently as possible, i.e. the agent must effectively leverage its sequentially observed aerial views when searching for the goal. To address these challenges, we propose GOMAA-Geo - a goal modality agnostic active geo-localization agent - for zeroshot generalization between different goal modalities. Our approach combines cross-modality contrastive learning to align representations across modalities with supervised foundation model pretraining and reinforcement learning to obtain highly effective navigation and localization policies. Through extensive evaluations, we show that GOMAA-Geo outperforms alternative learnable approaches and that it generalizes across datasets - e.g., to disaster-hit areas without seeing a single disaster scenario during training - and goal modalities - e.g., to ground-level imagery or textual descriptions, despite only being trained with goals specified as aerial views. Code and models will be made publicly available at this link.

Synthetic tabular data generation with differential privacy is a crucial problem to enable data sharing with formal privacy. Despite a rich history of methodological research and development, developing differentially private tabular data generators that can provide realistic synthetic datasets remains challenging. This paper introduces DP-LLMTGen -- a novel framework for differentially private tabular data synthesis that leverages pretrained large language models (LLMs). DP-LLMTGen models sensitive datasets using a two-stage fine-tuning procedure with a novel loss function specifically designed for tabular data. Subsequently, it generates synthetic data through sampling the fine-tuned LLMs. Our empirical evaluation demonstrates that DP-LLMTGen outperforms a variety of existing mechanisms across multiple datasets and privacy settings. Additionally, we conduct an ablation study and several experimental analyses to deepen our understanding of LLMs in addressing this important problem. Finally, we highlight the controllable generation ability of DP-LLMTGen through a fairness-constrained generation setting.

As Large Language Models (LLMs) are increasingly deployed in real-world settings, understanding the knowledge they implicitly use when making decisions is critical. One way to capture this knowledge is in the form of Bayesian prior distributions. We develop a prompt-based workflow for eliciting prior distributions from LLMs. Our approach is based on iterated learning, a Markov chain Monte Carlo method in which successive inferences are chained in a way that supports sampling from the prior distribution. We validated our method in settings where iterated learning has previously been used to estimate the priors of human participants -- causal learning, proportion estimation, and predicting everyday quantities. We found that priors elicited from GPT-4 qualitatively align with human priors in these settings. We then used the same method to elicit priors from GPT-4 for a variety of speculative events, such as the timing of the development of superhuman AI.

In the era of large language models, the demand for efficient use of computational resources has become critically important. Although parameter-efficient fine-tuning techniques have achieved results comparable to full fine-tuning, their application during the pre-training phase poses significant challenges. Specifically, employing parameter-efficient strategies at the onset of pre-training can severely compromise efficiency, especially in larger models. In this paper, building upon the fine-tuning method LoRA, we introduce a novel parameter-efficient training technique that frequently alters trainable part of parameters, facilitating effective pre-training. Our method not only achieves memory reductions and computational overhead comparable to current state-of-the-art parameter-efficient algorithms during the pre-training phase but also maintains accuracy levels comparable to those of full pre-training. We provide both theoretical analyses and empirical evidence to demonstrate the effectiveness of our approach.