arXiv subjects

Wentian Wang

Publications and source records attributed to Wentian Wang.

8 recordsLinked to original sources

arXiv2026

Interpretable GOHR Agents via Sparse Autoencoders

A central challenge in interpreting learned decision-making systems is to determine whether their internal representations contain concepts that help explain their behavior. We report interpretability experiments for a tokenized autoregressive Transformer agent in the Game of Hidden Rules (GOHR). We focus on a compact two-rule task in which both hidden rules map object shapes to target buckets, but with different permutations. The policy is trained on episodes sampled from these two hidden rules and then evaluated with fixed weights. It is never given a rule label and does not use an explicit rule classifier; any rule information must be inferred implicitly from interaction history. In this setting, the correct rule is not identifiable before the agent tries an informative move and observes accept/reject feedback. Sparse autoencoders (SAEs) trained on the agent's decision-token embeddings recover this structure. When held-out decisions are labeled by simple concepts such as the chosen shape or bucket, SAE dimensions that are highly selective for a concept cover most decisions where that concept is present. Individual SAE dimensions also correspond to interpretable strategies such as probing one rule hypothesis and switching after negative feedback.

cs.LG↗

arXiv2026

AI Learning and Conceptual Transfer in the Game of Hidden Rules

This report summarizes the work conducted on the Game of Hidden Rules (GOHR), focusing on reinforcement learning agents trained to infer hidden rules from trial-and-error feedback, representation design, rule difficulty analysis, transfer learning, generalization, and pseudo-bot-assisted human learning analysis. The report focuses on the Transformer-based A2C framework, Feature-Centric and Object-Centric representations, experimental findings, and classification of human learning data.

cs.AI↗

arXiv2026

CoRMA: Contrastive RMA for Contact-Rich Meta-Adaptation

We present CoRMA(Contrastive Robotic Motor Adaptation), a context-based meta-adaptation framework that modifies RMA for force-dominant assembly. CoRMA replaces raw simulator-parameter adaptation with a compact 6D simulator-only semantic contact context describing contact onset, lateral engagement, guided transition, contact direction, and jamming. A deployable causal Transformer adapter infers this context online from force, proprioceptive, and action histories using semantic regression and a force-regime contrastive objective. At deployment, oracle context is removed and replaced by the inferred context, enabling within-episode adaptation without demonstrations, privileged inputs, or gradient updates. We evaluate CoRMA on PegInsert, GearMesh, and NutThread in Isaac Lab / Isaac Sim 5.0 and on a real Marvin arm. Compared with FORGE baselines that achieve high simulation success but degrade substantially on hardware, CoRMA retains higher verified real success under controlled target-pose noise. These results support semantic contact inference as a reusable adaptation interface within a related assembly task family, while broader unseen-task generalization and Real2Sim calibration remain future work.

cs.RO↗

arXiv2025

Toward a Metrology for Artificial Intelligence: Hidden-Rule Environments and Reinforcement Learning

We investigate reinforcement learning in the Game Of Hidden Rules (GOHR) environment, a complex puzzle in which an agent must infer and execute hidden rules to clear a 6$\times$6 board by placing game pieces into buckets. We explore two state representation strategies, namely Feature-Centric (FC) and Object-Centric (OC), and employ a Transformer-based Advantage Actor-Critic (A2C) algorithm for training. The agent has access only to partial observations and must simultaneously infer the governing rule and learn the optimal policy through experience. We evaluate our models across multiple rule-based and trial-list-based experimental setups, analyzing transfer effects and the impact of representation on learning efficiency.

cs.LG↗

arXiv2025

Shock induced ignition and transition to detonation in the presence of mechanically induced non-linear acoustic forcing

We address the problem of shock induced ignition and transition to detonation in a reactive medium in the presence of mechanically induced fluctuations by a moving oscillating piston. For the inert problem prior to ignition, we provide a novel closed form model in Lagrangian coordinates for the generation of the train of compression and expansions, their steepening into a train of N-shock waves and their reflection on the lead shock, as well as the distribution energy dissipation rate in the induction zone. The model is found in excellent agreement with numerics. Reactive calculations were performed for hydrogen and ethylene fuels using a novel high-fidelity scheme to solve the reactive Euler equations written in Lagrangian coordinates. Different regimes of ignition and transition to detonation, controlled by the time scale of the forcing and the two time scales of the chemistry: the induction and reaction times. Two novel hot spot cascade mechanisms were identified. The first relies on the coherence between the sequence of hot spot formation set by the piston forcing and forward wave interaction with the lead shock, generalizing the classic runaway in fast flames. The second hot spot cascade is triggered by the feedback between the pressure pulse generated by the first generation hot spot cascade and the shock. For slow forcing, the sensitization is through a modification to the classic run-away process, while the high frequency regime leads to very localized sub-critical hot-spot formation controlled by the cumulative energy dissipation of the first generation shocks at a distance comparable to the shock formation location.

physics.flu-dyn↗

arXiv2024

MMLU-SR: A Benchmark for Stress-Testing Reasoning Capability of Large Language Models

We propose MMLU-SR, a novel dataset designed to measure the true comprehension abilities of Large Language Models (LLMs) by challenging their performance in question-answering tasks with modified terms. We reasoned that an agent that "truly" understands a concept can still evaluate it when key terms are replaced by suitably defined alternate terms, and sought to differentiate such comprehension from mere text replacement. In our study, we modified standardized test questions by replacing a key term with a dummy word along with its definition. The key term could be in the context of questions, answers, or both questions and answers. Notwithstanding the high scores achieved by recent popular LLMs on the MMLU leaderboard, we found a substantial reduction in model performance after such replacement, suggesting poor comprehension. This new benchmark provides a rigorous benchmark for testing true model comprehension, and poses a challenge to the broader scientific community.

cs.CL↗

arXiv2024

Detonation attenuation and quenching in hydrogen mixtures after the interaction with cylinders

The attenuation and quenching of H$_2$/O$_2$ detonations transmitted across a column of cylinders were studied experimentally and analytically at sub-atmospheric pressures. Two distinct transmission regimes were observed: successful transmission and complete quenching. The transition between the two regimes was found to correlate with the ratio of inter-cylinder separation distance (b) to a characteristic detonation scale for large blockage ratios (BRs), with critical limits comparable with those previously reported for detonation diffraction from slots. Based on available cell size measurements, the critical transmission limit was $b/λ=4.5\pm3$. The proposed theoretical model based on Whitham's geometric shock dynamics confirmed the equivalence between the detonation diffraction at abrupt area changes and around cylinders with large BRs. Complete quenching observed experimentally was accounted for by the weak transmitted shock strength upon detonation failure. For the tested BRs, the transmitted shock speed ranged between 50% and 60% of the Chapman-Jouguet detonation speed. As a result, the post-shock temperatures fell below the cross-over regime for hydrogen ignition, leading to very long ignition delay times ($t_i$) despite the temperature rise induced by Mach reflection. The long $t_i$ suppressed auto-ignition, while the high isentropic exponent prevented further convective mixing required for re-initiation. This explained the fundamental difference between arresting hydrogen and hydrocarbon detonations, for which transmitted fast flames punctuated by auto-ignition events were always observed. The transmitted shock strength was found to be well-predicted by our previous self-similar multiple discontinuity gas-dynamic model. Over a very narrow range near the critical conditions, both hot spot re-ignition and detonation re-initiation from Mach shock reflections were observed.

physics.flu-dyn↗

arXiv2015

Chapman-Jouguet deflagrations and their transition to detonation

We study experimentally fast flames and their transition to detonation in mixtures of methane, ethane, ethylene, acetylene, and propane mixtures with oxygen. Following the interaction of a detonation wave with a column of cylinders of varying blockage ratio, the experiments demonstrate that the fast flames established are Chapman-Jouguet deflagrations, in excellent agreement with the self-similar model of Radulescu et al. (2015). The experiments indicate that these Chapman-Jouguet deflagrations dynamically restructure and amplify into fewer stronger modes until the eventual transition to detonation. The transition length to a self-sustained detonation was found to correlate very well with the mixtures' sensitivity to temperature fluctuations, reflected by the $χ$ parameter introduced by Radulescu, which is the product of the non-dimensional activation energy $E_a/RT$ and the ratio of chemical induction to reaction time $t_i/t_r$. Correlation of the measured DDT lengths determined that the relevant characteristic time scale from chemical kinetics controlling DDT is the energy release or excitation time $t_r$. Correlations with the cell size also capture the dependence of the DDT length on $χ$ for fixed blockage ratios.

physics.flu-dyn↗