In Korea, Nuclear Power Plants continually prepare and revise the final safety analysis report (FSAR), which is essential for predicting the source-term information of radioactive materials throughout their lifecycle. This report provides detailed information on the source-terms of Chalk River Unidentified Deposits (CRUD). The management of radioactive corrosion products, present on nuclear fuel surfaces and within the primary system, is crucial for ensuring equipment integrity and controlling worker exposure during operation. Unlike earlier studies that focused solely on the radioactive inventory of major components, this study established an integrated evaluation framework by coupling the CRUDTRAN, ORIGEN2, and PHITS codes to quantitatively assess the relative contribution of the CRUD source-term on the core support barrel (CSB) to worker doses, in comparison with the activation source-terms of the reactor pressure vessel (RPV). By calculating these specific source-terms and conducting subsequent radiation exposure assessments, this work underscores the critical significance of the CRUD source-term during the decommissioning phase. The comparative dose assessment revealed that in the absence of system decontamination, the radiation exposure contributed by the CRUD source-term exceeded that from the structural activation of the RPV across the range of deposition-fraction assumptions, from about 1.5 times under the minimum assumption to about 117 times under the maximum assumption. These quantitative findings indicate that CRUD is a major source of worker exposure during the decommissioning phase, which should not be neglected relative to the activation of internal components.
Vision-Language-Action (VLA) models have recently achieved notable progress in end-to-end autonomous driving by integrating perception, reasoning, and control within a unified multimodal framework. However, they often lack explicit modeling of temporal dynamics and global world consistency, which limits their foresight and safety. In contrast, world models can simulate plausible future scenes but generally struggle to reason about or evaluate the imagined future they generate. In this work, we present VLA-World, a simple yet effective VLA world model that unifies predictive imagination with reflective reasoning to improve driving foresight. VLA-World first uses an action-derived feasible trajectory to guide the generation of the next-frame image, capturing rich spatial and temporal cues that describe how the surrounding environment evolves. The model then reasons over this self-generated future imagined frame to refine the predicted trajectory, achieving higher performance and better interpretability. To support this pipeline, we curate nuScenes-GR-20K, a generative reasoning dataset derived from nuScenes, and employ a three-stage training strategy that includes pretraining, supervised fine-tuning, and reinforcement learning. Extensive experiments demonstrate that VLA-World consistently surpasses state-of-the-art VLA and world-model baselines on both planning and future-generation benchmarks. Project page: https://vlaworld.github.io
We develop a unified mathematical framework for certified Top-k attention truncation that quantifies approximation error at both the distribution and output levels. For a single attention distribution P and its Top-k truncation P̂, we show that the total-variation distance coincides with the discarded softmax tail mass and satisfies TV(P,P̂)=1-e^-KL(P̂‖ P), yielding sharp Top-k-specific bounds in place of generic inequalities. From this we derive non-asymptotic deterministic bounds – from a single boundary gap through multi-gap and blockwise variants – that control TV(P,P̂) using only the ordered logits. Using an exact head-tail decomposition, we prove that the output error factorizes as Attn(q,K,V)-Attn_k(q,K,V)_2=τμ_tail-μ_head_2 with τ=TV(P,P̂), yielding a new head-tail diameter bound Attn(q,K,V)-Attn_k(q,K,V)_2 diam_H,T and refinements linking the error to Var_P(V). Under an i.i.d. Gaussian score model s_i∼𝒩(μ,σ^2) we derive closed-form tail masses and an asymptotic rule for the minimal k_ε ensuring TV(P,P̂)≤ε, namely k_ε/n_c(σ+Φ^-1(ε)). Experiments on bert-base-uncased and synthetic logits confirm the predicted scaling of k_ε/n and show that certified Top-k can reduce scored keys by 2-4× on average while meeting the prescribed total-variation budget.
A series of Na/Fe and Na/Mn/Fe catalysts were synthesized and applied in Fischer-Tropsch Synthesis (FTS) to investigate the promotion mechanisms of Na, Mn and their synergy. Na increases the particle size of Fe2O3 and inhibits its reduction due to the alkaline nature. In contrast, Mn decreases particle size and enhances the catalyst's specific surface area. Mossbauer spectroscopy analysis reveals that both post-reaction Na/Fe and Na/Mn/ Fe contained exclusively chi-Fe5C2 as the iron carbide phase, with no detectable Fe3O4. FTS performance shows that Na/Fe catalysts exhibited high selectivity for long-chain hydrocarbons and high C2-C4 olefin/paraffin (O/P) ratio, but suffers from poor stability with a significant decline in CO conversion over time. When the Na content is relatively high, the FTS performance of the Na/Mn/Fe catalysts is similar to that of the Na/Fe catalysts. It is noteworthy that the optimal synergistic Na-Mn interaction enables the 1Na/5Mn/10Fe catalyst to exhibit high long-chain hydrocarbon selectivity, favorable C2-C4 O/P ratio and good stability. The differing performance of Na/Mn/Fe catalysts with variable Na content indicates competing influences within the Na-Mn interaction mechanism. Detailed comparative analysis of the post-reaction phases, specific surface areas and surface deposited carbon of the spent 1Na/10Fe and 1Na/5Mn/10Fe catalysts indicate that both contain exclusively chi-Fe5C2 and exhibit similar specific surface areas. Thermogravimetric analysis (TG) reveals reduced deposited carbon content on the surface of spent 1Na/5Mn/10Fe catalyst compared to 1Na/10Fe, which is the crucial factor contributing to the stability. TEM and elemental mapping reveal that MnOx species are distributed around Fe species in the 1Na/5Mn/10Fe catalyst, forming a quasi-core-shell structure. This MnOx shell inhibits carbon deposition, which accounts for the enhanced stability of the 1Na/5Mn/10Fe catalyst.
Post-training of large language models routinely interleaves supervised fine-tuning (SFT) with reinforcement learning (RL). These two methods have different objectives: SFT minimizes the cross-entropy loss between model outputs and expert responses, while RL maximizes reward signals derived from human preferences or rule-based verifiers. Modern reasoning models have widely adopted the practice of alternating SFT and RL training. However, there is no theoretical account of whether they can be decoupled. We prove that decoupling is impossible in either order: (1) SFT-then-RL coupling: RL increases SFT loss under SFT optimality and (2) RL-then-SFT coupling: SFT lowers the reward achieved by RL. Experiments on Qwen3-0.6B confirm the predicted degradation, verifying that SFT and RL cannot be separated without loss of prior performance in the post-training
