Commonsense Reasoning

Throughout life, we accumulate knowledge of how typical certain events are. We use this knowledge in daily conversation, often interpreting meaning beyond what has been said explicitly. In this work, we take steps to better understand this meaning recovery, investigating both atypicality and typicality inferences. First, we construct two corpora, EventTyp and the human-validated EventTypGold. We conduct large-scale human and LLM experiments and find that for humans, three CUES predict inferences about typicality: Commonsense Knowledge, Utterance Emphasis, and Snark. However, in most experimental settings, we find that LLMs lack the prerequisite Commonsense Knowledge necessary to accurately assess an event's general typicality, and therefore struggle to follow these CUES.

Dynamic interactive object search in large-scale human environments presents substantial challenges for existing methods. Current scene representations like 3D Scene Graphs (3DSG) only provide coarse-grained spatial segmentation and cannot identify functional areas such as storage or leisure areas. Without functional area understanding, existing methods are constrained to exhaustive sequential exploration at large scales, resulting in inefficient search behaviors -- particularly in open-layout environments with numerous interactive objects such as drawers and cabinets. Moreover, these methods lack adaptability to environmental dynamics such as object relocations. To address these limitations, this paper proposes CMAR-search, a novel framework built upon Commonsense and Memory Augmented Reasoning (CMAR). Our approach leverages commonsense about area functionalities and aggregates environmental memory to construct a Functional 3D Scene Graph (F3DSG), which organizes the environment into functional areas with their associated containers. Through this structured representation, CMAR enables hierarchical action planning at both macro-area and micro-container levels, empowering the system to efficiently identify and inspect semantically relevant areas for effective object search. Notably, CMAR continuously integrates real-time perception, accumulated memory, and commonsense to dynamically relocalize objects in changing environments. Extensive experiments in simulation and real-world settings demonstrate that CMAR-search significantly surpasses state-of-the-art baselines in both success rate and search efficiency for object search in dynamic interactive environments.

Compositional generalization remains challenging when language models must combine familiar reasoning operations in unfamiliar ways. The Scenario-Based Commonsense Reasoning Evaluation (SCoRE) 2026 tests this ability on three mixed domains absent from training and requires models to identify the complete set of correct options for each question. We introduce Route-Verify-Vote (RVV), a framework for procedure-conditioned self-consistency that uses language models without parameter updates. Route uses the provided domain label to select a reasoning procedure that guides the model in representing and applying the relevant constraints. Verify prompts the model to assess each option against those constraints. Vote aggregates complete answer sets and allocates additional samples to questions with a small vote-count margin between the two most frequent sets. Samples for each question follow the same domain-specific procedure. On the official test set, voting over 16 sampled answer sets per question achieves an exact-set accuracy of 74.6%. Adaptive RVV reaches 77.3%, and combining models on selected domain routes raises accuracy to 79.4%. The final system ranked second among participating systems. These results support domain-specific reasoning procedures and answer-set disagreement as useful tools for allocating inference-time computation in mixed-domain reasoning.

Embodied Vision-Language-Action (VLA) models are typically obtained by fine-tuning powerful pretrained VLMs on robotics data, yet it is unclear how much commonsense and factual knowledge they retain after adaptation. Failures on knowledge-sensitive tasks are ambiguous, conflating missing knowledge with poor generalization of low-level control. We introduce Act2Answer, a lightweight protocol that adapts VLM knowledge benchmarks to VLA evaluation by requiring agents to answer through action. Each question becomes a short tabletop episode where the agent performs a single object-placement action to select among candidate answers, yielding an action-grounded success rate with reduced control confounds. We curate a test suite of such environments across diverse commonsense and world-knowledge categories and introduce layerwise intent probing to localize answer-relevant information across the VLM backbone and action head. In a large-scale study of 7 VLA models and 9 VLM baselines, we systematically rank models across categories, finding that VLAs show solid performance on simple concepts while exhibiting larger gaps on richer semantic categories relative to their source VLMs, that VQA co-training is associated with better knowledge retention, and that answer-relevant signals peak in middle VLA layers but attenuate in upper layers. Act2Answer is available at https://tttonyalpha.github.io/act2answer/.

When large language models (LLMs) fail to generalize or make content-sensitive errors in reasoning, it is often taken as evidence that LLMs are not truly reasoning, but rather performing a kind of pattern matching. The implication is that human behavior does not exhibit the same types of failures because human reasoning relies on principled and content-invariant world models. We test this assumption by first evaluating humans and LLMs on their ability to engage in common-sense reasoning about a variety of everyday situations. Our results reveal convergent patterns of reasoning across 46 LLMs and two cohorts of human participants. We then ask whether this behavioral convergence is due to LLMs having acquired content-invariant world models or a set of pattern-matching heuristics by characterizing the roles of content-invariant and content-sensitive model neurons in producing human-like responses. We find that while LLMs encode both content-invariant and content-sensitive representations, it is content-sensitive mechanisms which are causally responsible for aligning models with humans. Taken together, our results suggest that everyday causal reasoning in people and LLMs makes heavy use of pattern-matching.

Learning-driven Scene Graph Generation (SGG) models excel on frequent relation types but degrade sharply under annotation sparsity, failing to capture reliable visual commonsense knowledge. We propose a model-agnostic, semantically-guided knowledge refinement framework that systematically mines commonsense-grounded constraints from training data - capturing spatial, functional, and qualitative relational regularities - and uses general declarative commonsense reasoning to correct and refine ranked SGG predictions at inference time. The framework requires no manual rule authoring, no model retraining, and transfers across datasets and architectures. On three standard benchmarks, we obtain consistent improvements over strong baselines, demonstrating that structured visual commonsense reasoning over deep scene semantics is a practical and effective complement to purely learning-based scene graph generation.

Parameter-efficient fine-tuning (PEFT) has become a practical route for adapting large language models to downstream tasks, with LoRA-style methods being particularly attractive because they are inexpensive to train and easy to deploy. Most LoRA variants, however, revise the update rule within the weight space of each layer and leave the intermediate representations formed by deeper layers largely unused. We propose Echo-LoRA, a cross-layer representation injection method for parameter-efficient fine-tuning. During training, Echo-LoRA collects boundary hidden states from deeper source layers, aggregates them into a sample-level echo representation, and uses lightweight projection and gating networks to inject the resulting signal into shallow LoRA or DoRA modules. Answer-only masking, masked distillation, and stochastic routing are used to keep this auxiliary path stable and to reduce the gap between training and inference. On eight commonsense reasoning benchmarks, Echo-LoRA exceeds the reported LoRA baselines by 5.7 percentage points on average across LLaMA-7B, LLaMA2-7B, and LLaMA3-8B. Under reproduced LoRA baselines in our unified implementation, the average gain is 3.0 points; when combined with DoRA, the gain is 2.7 points. The Echo path is discarded after training, so the deployed model keeps the original low-rank LoRA/DoRA form and adds neither inference-time parameters nor inference computation.

Masked diffusion models (MDM) exhibit superior generalization when learned using a Partial masking scheme (Prime). This approach converts tokens into sub-tokens and models the diffusion process at the sub-token level. We identify two limitations of the MDM-Prime framework. First, we find that the functional form of the subtokenizer significantly increases the cross-entropy loss in the objective when paired with commonly used Byte-Pair-Encoding (BPE) tokenizers. Second, we lack tools to guide the hyperparameter choice of the token granularity in the subtokenizer. To address these limitations, we analyze the optimal design of the subtokenizer that minimizes MDM-Prime training objective and develop MDM-Prime-v2, a masked diffusion language model which incorporates Binary Encoding and Index Shuffling. Our analysis characterizes how token granularity and sub-token entropy influence the training objective and downstream performance, providing principled criteria for subtokenizer design. When extending the model size to 1.1B parameters, MDM-Prime-v2 demonstrates superior average zero-shot accuracy across eight commonsense reasoning benchmarks, outperforming similar-sized baselines including GPT-Neo, OPT, Pythia, Bloom, SMDM, and TinyLLaMA.

Large language models (LLMs) achieve impressive scores on standard benchmarks yet routinely fail questions that any human would answer correctly in seconds. We introduce BrainBench, a benchmark of 100 brainteaser questions spanning 20 carefully designed categories, each targeting a specific commonsense reasoning failure mode in LLMs. Categories range from implicit physical constraints ("Should I walk or drive my rental car to the return lot?") to semantic scope tricks and default assumption hijacks. We evaluate eight frontier models -- four from the Claude family and four from the GPT family -- using a zero-shot protocol with 10 independent runs per question. The best model, Claude Opus 4.6 with extended thinking, achieves only 80.3% accuracy; the worst, GPT-4o, scores 39.7%. Even top-performing models exhibit a 6-16 percentage-point gap between accuracy and consistency, revealing stochastic reasoning. Cross-lingual evaluation in Chinese shows most models degrade by 2-8 percentage points, confirming that these failures reflect reasoning deficits rather than language-specific artifacts. BrainBench provides a fine-grained diagnostic tool for identifying where and why LLMs substitute surface heuristics for genuine commonsense reasoning.

We present MMCOMET, the first multimodal commonsense knowledge graph (MMKG) that integrates physical, social, and eventive knowledge. MMCOMET extends the ATOMIC2020 knowledge graph to include a visual dimension, through an efficient image retrieval process, resulting in over 900K multimodal triples. This new resource addresses a major limitation of existing MMKGs in supporting complex reasoning tasks like image captioning and storytelling. Through a standard visual storytelling experiment, we show that our holistic approach enables the generation of richer, coherent, and contextually grounded stories than those produced using text-only knowledge. This resource establishes a new foundation for multimodal commonsense reasoning and narrative generation.

Commonsense reasoning is a key language model capability, as it is purportedly a prerequisite for many basic tasks, unlike specific factual knowledge. It is often measured with multiple-choice questions (MCQ) benchmarks, e.g. HellaSwag and PIQA. Some of these benchmarks, however, are outdated and contain numerous validity issues. We illustrate some typical validity issues with a case study on HellaSwag, one of the most popular and problematic benchmarks for commonsense reasoning. The issues we find range from basic ungrammaticality and numerous typos to misleading prompts or equally correct options. We show that if we remove question prompts or replace them with "Lorem ipsum dolor...", about 68% of model predictions do not change. We argue that this occurs due to inner flaws in the benchmark, not mere contamination that might be present in some models. Since benchmark scores are an essential part of model selection in both research and commercial applications, these issues can have severe consequences. Based on our findings, we propose BenCheck, a package for benchmark validity analysis that encapsulates the main checks performed in our case study and can be used to audit commonsense reasoning benchmarks. We apply these checks to PIQA, Global PIQA, and Winogrande.