We introduce TESS 2, a general instruction-following diffusion language model that outperforms contemporary instruction-tuned diffusion models, as well as matches and sometimes exceeds strong autoregressive (AR) models. We train TESS 2 by first adapting a strong AR model via continued pretraining with the usual cross-entropy as diffusion loss, and then performing further instruction tuning. We find that adaptation training as well as the choice of the base model is crucial for training good instruction-following diffusion models. We further propose reward guidance, a novel and modular inference-time guidance procedure to align model outputs without needing to train the underlying model. Finally, we show that TESS 2 further improves with increased inference-time compute, highlighting the utility of diffusion LMs in having fine-grained controllability over the amount of compute used at inference time. Code and models are available at https://github.com/hamishivi/tess-2.

"Socrates is human. All humans are mortal. Therefore, Socrates is mortal." This classical example demonstrates two-hop reasoning, where a conclusion logically follows from two connected premises. While transformer-based Large Language Models (LLMs) can make two-hop reasoning, they tend to collapse to random guessing when faced with distracting premises. To understand the underlying mechanism, we train a three-layer transformer on synthetic two-hop reasoning tasks. The training dynamics show two stages: a slow learning phase, where the 3-layer transformer performs random guessing like LLMs, followed by an abrupt phase transitions, where the 3-layer transformer suddenly reaches $100%$ accuracy. Through reverse engineering, we explain the inner mechanisms for how models learn to randomly guess between distractions initially, and how they learn to ignore distractions eventually. We further propose a three-parameter model that supports the causal claims for the mechanisms to the training dynamics of the transformer. Finally, experiments on LLMs suggest that the discovered mechanisms generalize across scales. Our methodologies provide new perspectives for scientific understandings of LLMs and our findings provide new insights into how reasoning emerges during training.

Current image captioning systems lack the ability to link descriptive text to specific visual elements, making their outputs difficult to verify. While recent approaches offer some grounding capabilities, they cannot track object identities across multiple references or ground both actions and objects simultaneously. We propose a novel ID-based grounding system that enables consistent object reference tracking and action-object linking, and present GroundCap, a dataset containing 52,016 images from 77 movies, with 344 human-annotated and 52,016 automatically generated captions. Each caption is grounded on detected objects (132 classes) and actions (51 classes) using a tag system that maintains object identity while linking actions to the corresponding objects. Our approach features persistent object IDs for reference tracking, explicit action-object linking, and segmentation of background elements through K-means clustering. Human evaluation demonstrates our approach's effectiveness in producing verifiable descriptions with coherent object references. Introduction One of the primary goals combining computer vision and natural language processing is to enable machines to understand and communicate about visual scenes. This objective encompasses numerous tasks, including recognizing objects, describing their attributes and relationships, and providing contextually relevant descriptions of scenes [1]. While significant progress has been made in image classification, object detection, and image captioning, a critical aspect of human visual communication remains under-explored: the ability to ground language to specific elements within an image. Consider a scenario where two people are discussing a crowded street scene. One might say, "Look at that car." to which the other might respond, "Which one?". The first person would likely point to the specific car they're referring to while simultaneously describing it with more detail.

This paper presents DataSciBench, a comprehensive benchmark for evaluating Large Language Model (LLM) capabilities in data science. Recent related benchmarks have primarily focused on single tasks, easily obtainable ground truth, and straightforward evaluation metrics, which limits the scope of tasks that can be evaluated. In contrast, DataSciBench is constructed based on a more comprehensive and curated collection of natural and challenging prompts for uncertain ground truth and evaluation metrics. We develop a semi-automated pipeline for generating ground truth (GT) and validating evaluation metrics. This pipeline utilizes and implements an LLM-based self-consistency and human verification strategy to produce accurate GT by leveraging collected prompts, predefined task types, and aggregate functions (metrics). Furthermore, we propose an innovative Task - Function - Code (TFC) framework to assess each code execution outcome based on precisely defined metrics and programmatic rules. Our experimental framework involves testing 6 API-based models, 8 open-source general models, and 9 open-source code generation models using the diverse set of prompts we have gathered. This approach aims to provide a more comprehensive and rigorous evaluation of LLMs in data science, revealing their strengths and weaknesses. Experimental results demonstrate that API-based models outperform open-sourced models on all metrics and Deepseek-Coder-33B-Instruct achieves the highest score among open-sourced models. We release all code and data at https://github.com/THUDM/DataSciBench.

Navigating unfamiliar environments presents significant challenges for household robots, requiring the ability to recognize and reason about novel decoration and layout. Existing reinforcement learning methods cannot be directly transferred to new environments, as they typically rely on extensive mapping and exploration, leading to time-consuming and inefficient. To address these challenges, we try to transfer the logical knowledge and the generalization ability of pre-trained foundation models to zero-shot navigation. By integrating a large vision-language model with a diffusion network, our approach named \mname ~constructs a visual predictor that continuously predicts the agent's potential observations in the next step which can assist robots generate robust actions. Furthermore, to adapt the temporal property of navigation, we introduce temporal historical information to ensure that the predicted image is aligned with the navigation scene. We then carefully designed an information fusion framework that embeds the predicted future frames as guidance into goal-reaching policy to solve downstream image navigation tasks. This approach enhances navigation control and generalization across both simulated and real-world environments. Through extensive experimentation, we demonstrate the robustness and versatility of our method, showcasing its potential to improve the efficiency and effectiveness of robotic navigation in diverse settings.

Conversational Product Search (CPS) is confined to simulated conversations due to the lack of real-world CPS datasets that reflect human-like language. Additionally, current conversational datasets are limited to support cross-market and multi-lingual usage. In this paper, we introduce a new CPS data collection protocol and present PSCon, a novel CPS dataset designed to assist product search via human-like conversations. The dataset is constructed using a coached human-to-human data collection protocol and supports two languages and dual markets. Also, the dataset enables thorough exploration of six subtasks of CPS: user intent detection, keyword extraction, system action prediction, question selection, item ranking, and response generation. Furthermore, we also offer an analysis of the dataset and propose a benchmark model on the proposed CPS dataset.

Large language models (LLMs) have revolutionized machine learning due to their ability to capture complex interactions between input features. Popular post-hoc explanation methods like SHAP provide marginal feature attributions, while their extensions to interaction importances only scale to small input lengths ($\approx 20$). We propose Spectral Explainer (SPEX), a model-agnostic interaction attribution algorithm that efficiently scales to large input lengths ($\approx 1000)$. SPEX exploits underlying natural sparsity among interactions -- common in real-world data -- and applies a sparse Fourier transform using a channel decoding algorithm to efficiently identify important interactions. We perform experiments across three difficult long-context datasets that require LLMs to utilize interactions between inputs to complete the task. For large inputs, SPEX outperforms marginal attribution methods by up to 20% in terms of faithfully reconstructing LLM outputs. Further, SPEX successfully identifies key features and interactions that strongly influence model output. For one of our datasets, HotpotQA, SPEX provides interactions that align with human annotations. Finally, we use our model-agnostic approach to generate explanations to demonstrate abstract reasoning in closed-source LLMs (GPT-4o mini) and compositional reasoning in vision-language models.

We introduce Faina, the first dataset for fallacy detection that embraces multiple plausible answers and natural disagreement. Faina includes over 11K span-level annotations with overlaps across 20 fallacy types on social media posts in Italian about migration, climate change, and public health given by two expert annotators. Through an extensive annotation study that allowed discussion over multiple rounds, we minimize annotation errors whilst keeping signals of human label variation. Moreover, we devise a framework that goes beyond "single ground truth" evaluation and simultaneously accounts for multiple (equally reliable) test sets and the peculiarities of the task, i.e., partial span matches, overlaps, and the varying severity of labeling errors. Our experiments across four fallacy detection setups show that multi-task and multi-label transformer-based approaches are strong baselines across all settings. We release our data, code, and annotation guidelines to foster research on fallacy detection and human label variation more broadly.

Retrieval-Augmented Generation (RAG) systems have shown substantial benefits in applications such as question answering and multi-turn dialogue [22]. However, traditional RAG methods, while leveraging static knowledge bases, often overlook the potential of dynamic historical information in ongoing conversations. To bridge this gap, we introduce DH-RAG, a Dynamic Historical Context-Powered Retrieval-Augmented Generation Method for Multi-Turn Dialogue. DH-RAG is inspired by human cognitive processes that utilize both long-term memory and immediate historical context in conversational responses [32]. DH-RAG is structured around two principal components: a History-Learning based Query Reconstruction Module, designed to generate effective queries by synthesizing current and prior interactions, and a Dynamic History Information Updating Module, which continually refreshes historical context throughout the dialogue. The center of DH-RAG is a Dynamic Historical Information database, which is further refined by three strategies within the Query Reconstruction Module: Historical Query Clustering, Hierarchical Matching, and Chain of Thought Tracking. Experimental evaluations show that DH-RAG significantly surpasses conventional models on several benchmarks, enhancing response relevance, coherence, and dialogue quality.

Current recommendation systems powered by large language models (LLMs) often underutilize their reasoning capabilities due to a lack of explicit logical structuring. To address this limitation, we introduce CoT-Rec, a framework that integrates Chain-of-Thought (CoT) reasoning into LLM-driven recommendations by incorporating two crucial processes: user preference analysis and item perception evaluation. CoT-Rec operates in two key phases: (1) personalized data extraction, where user preferences and item perceptions are identified, and (2) personalized data application, where this information is leveraged to refine recommendations. Our experimental analysis demonstrates that CoT-Rec improves recommendation accuracy by making better use of LLMs' reasoning potential. The implementation is publicly available at https://anonymous.4open.science/r/CoT-Rec.