Chain-of-Thought (CoT) reasoning is a critical capability for large language models (LLMs), enabling them to tackle com- plex multi-step tasks. While base LLMs, pre-trained on general text corpora, often struggle with reasoning due to a lack of specialized training, recent studies reveal their latent reason- ing potential tied to hidden states. However, existing hidden state manipulation methods, such as linear activation steering, suffer from limitations due to their rigid and unconstrained nature, often leading to distribution shifts and degraded text quality. In this work, we propose a novel approach for elic- iting CoT reasoning from base LLMs through hidden state manipulation grounded in probabilistic conditional generation. By reformulating the challenge as an optimization problem with a balanced likelihood and prior regularization framework, our method guides hidden states toward reasoning-oriented trajectories while preserving linguistic coherence. Extensive evaluations across mathematical, commonsense, and logical reasoning benchmarks demonstrate that our approach con- sistently outperforms existing steering methods, offering a theoretically principled and effective solution for enhancing reasoning capabilities in base LLMs.

Recent advancements in large language models (LLMs) have resulted in increasingly anthropomorphic language concerning the ability of LLMs to reason. Whether reasoning in LLMs should be understood to be inherently different is, however, widely debated. We propose utilizing a representation engineering approach wherein model activations are read from the residual stream of an LLM when processing a reasoning task. The activations are used to derive a control vector that is applied to the model as an inference-time intervention, modulating the representational space of the model, to improve performance on the specified task. We publish the code for deriving control vectors and analyzing model representations. The method allows us to improve performance on reasoning benchmarks and assess how control vectors influence the final logit distribution of a model via metrics such as KL divergence and entropy. We apply control vectors to Mistral-7B-Instruct and a range of Pythia models on an inductive, a deductive and mathematical reasoning task. We show that an LLM can, to a certain degree, be controlled to improve its perceived reasoning ability by modulating activations. The intervention is dependent upon the ability to reliably extract the model's typical state when correctly solving a task. Our results suggest that reasoning performance can be modulated in the same manner as other information-processing tasks performed by LLMs and demonstrate that we are capable of improving performance on specific tasks via a simple intervention on the residual stream with no additional training. Many recent developments in the study of artificial intelligence and more specifically large language models (LLMs) have focused on improving their ability to solve reasoning tasks.

We propose a machine learning approach to the optimal control of fluid restless multi-armed bandits (FRMABs) with state equations that are either affine or quadratic in the state variables. By deriving fundamental properties of FRMAB problems, we design an efficient machine learning based algorithm. Using this algorithm, we solve multiple instances with varying initial states to generate a comprehensive training set. We then learn a state feedback policy using Optimal Classification Trees with hyperplane splits (OCT-H). We test our approach on machine maintenance, epidemic control and fisheries control problems. Our method yields high-quality state feedback policies and achieves a speed-up of up to 26 million times compared to a direct numerical algorithm for fluid problems.

Large language models (LLMs), despite their breakthroughs on many challenging benchmark tasks, lean to generate verbose responses and lack the controllability of output complexity, which is usually preferred by human users in practice. In this paper, we study how to precisely control multiple linguistic complexities of LLM output by finetuning using off-the-shelf data. To this end, we propose multi-control tuning (MCTune), which includes multiple linguistic complexity values of ground-truth responses as controls in the input for instruction tuning. We finetune LLaMA2-7B on Alpaca-GPT4 and WizardLM datasets. Evaluations on widely used benchmarks demonstrate that our method does not only improve LLMs' multi-complexity controllability substantially but also retains or even enhances the quality of the responses as a side benefit.

Motion forecasting transforms sequences of past movements and environment context into future motion. Recent methods rely on learned representations, resulting in hidden states that are difficult to interpret. In this work, we use natural language to quantize motion features in a human-interpretable way, and measure the degree to which they are embedded in hidden states. Our experiments reveal that hidden states of motion sequences are arranged with respect to our discrete sets of motion features. Following these insights, we fit control vectors to motion features, which allow for controlling motion forecasts at inference. Consequently, our method enables controlling transformer-based motion forecasting models with textual inputs, providing a unique interface to interact with and understand these models. Our implementation is available at https://github.com/kit-mrt/future-motion

Prior work in style-controlled text generation has focused on tasks such as emulating the style of prolific literary authors, producing formal or informal text, and the degree of toxicity of generated text. Plentiful demonstrations of these styles are available, and as a result modern language models are often able to emulate them, either via prompting or discriminative control. However, in applications such as writing assistants, it is desirable for language models to produce text in an author-specific style on the basis of a small writing sample. We find that instruction-tuned language models can struggle to reproduce author-specific style demonstrated in a prompt. Instead, we propose to guide a language model to generate text in a target style using contrastively-trained representations that capture stylometric features. A central challenge in doing so is that an author's writing is characterized by surprising token choices under a generic language model. To reconcile this tension, we combine generative re-scoring to achieve an author-specific model, with discriminative control to ensure style consistency at the sequence-level. The combination of these approaches is found to be particularly effective at adhering to an author-specific style in a variety of conditions, including unconditional generation and style transfer, and is applicable to any underlying language model without requiring fine-tuning.

We propose a novel unsupervised method to autoregressively generate videos from a single frame and a sparse motion input. Our trained model can generate realistic object-to-object interactions and separate the dynamics and the extents of multiple objects despite only observing them under correlated motion activities. Key components in our method are the randomized conditioning scheme, the encoding of the input motion control, and the randomized and sparse sampling to break correlations. Our model, which we call YODA, has the ability to move objects without physically touching them. We show both qualitatively and quantitatively that YODA accurately follows the user control, while yielding a video quality that is on par with or better than state of the art video generation prior work on several datasets. For videos, visit our project website https://araachie.github.io/yoda.

Unmanned ground vehicles (UGVs) in unstructured environments mostly operate through teleoperation. To enable stable teleoperated driving in unstructured environments, some research has suggested driver assistance and evaluation methods that involve user studies, which can be costly and require lots of time and effort. A simulation model-based approach has been proposed to complement the user study; however, the models on teleoperated driving do not account for unstructured environments. Our proposed solution involves simulation models of teleoperated driving for drivers that utilize a deep generative model. Initially, we build a teleoperated driving simulator to imitate unstructured environments based on previous research and collect driving data from drivers. Then, we design and implement the simulation models based on a conditional variational autoencoder (CVAE). Our evaluation results demonstrate that the proposed teleoperated driving model can generate data by simulating the driver appropriately in unstructured canyon terrains.

Location-aware networks will introduce new services and applications for modern convenience, surveillance, and public safety. In this paper, we consider the problem of cooperative localization in a wireless network where the position of certain anchor nodes can be controlled. We introduce an active planning method that aims at moving the anchors such that the information gain of future measurements is maximized. In the control layer of the proposed method, control inputs are calculated by minimizing the traces of approximate inverse Bayesian Fisher information matrixes (FIMs). The estimation layer computes estimates of the agent states and provides Gaussian representations of marginal posteriors of agent positions to the control layer for approximate Bayesian FIM computations. Based on a cost function that accumulates Bayesian FIM contributions over a sliding window of discrete future timesteps, a receding horizon (RH) control is performed. Approximations that make it possible to solve the resulting tree-search problem efficiently are also discussed. A numerical case study demonstrates the intelligent behavior of a single controlled anchor in a 3-D scenario and the resulting significantly improved localization accuracy.

The conflict between increasing demand for computing and sluggish grow of hardware capability triggers the heated development of approximate computing, which has achieved massive success in both industry and research community. Many applications that do not require utterly accurate computation can achieve tremendous acceleration and drastic reduction of the energy consumption by leveraging approximate computing, especially in domains that call for real-time calculation, fast response and low power consumption such as learning [27], image processing [19] and scientific computation [24]. Approximation computing can be conduct in different hierarchies, such as hardware [6], [18], system and software levels. Various approximate computing architectures [17], [19], [27] are advocated. Neural network (NN) based approximate computing focus on the acceleration in software-level and has many advantages when compared to previous methods. First, neural networks are proved to be able to fit any continuous function [12], and thus this method can universally be adopted by different tasks. Second, enormous parallelism in the neural networks is exploited by the rapid advancement of various neural network accelerators.