Quality-Gated Length Advantage Shaping (QGLAS), which first computes advantages from quality rewards alone, then adds bounded bonuses only to shorter positive-advantage responses, leaving all other advantages unchanged, consistently achieves a stronger quality--length trade-off than representative baselines.
Zi-Jun Weng, Zhong-An Bi, Xuan-Ang Gao et al.· 0 citations
Reinforcement learning (RL) is increasingly used to post-train vision-language-action (VLA) models, yet how RL reshapes these policies remains poorly understood. We find that RL across widely used flow-based VLA models, including $\pi_{0.5}$ and GR00T~N1.5/N1.6, on LIBERO, ManiSkill, MetaWorld, and CALVIN induces subst...
Minjae Oh, Yoonah Park, Jongwon Lim et al.· 0 citations
Large Language Model (LLM) routers commonly rely on neural query embeddings, with larger encoders expected to better capture query intent and difficulty. Yet scaling Qwen2.5 encoders from 0.5B to 72B parameters brings little improvement in routing accuracy (Figure 1b), suggesting that small encoders may already capture...
Yi-Fan Lu, Qi-Yue Zhang, Haotian Shan et al.· 0 citations
PQ-HSA (hybrid sparse-approximate attention) attends the selected tokens with their original keys and values, and the unselected tokens, the background, enter the same softmax through those scores, summed per inverted list and multiplied by the list's mean value.
Kun-Ming Shao, Jie-Run Chen, Yan-Li Wang et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
In-context learning (ICL) is crucial for boosting the inference performance of large language models (LLMs). However, the effectiveness of ICL in LLMs is greatly influenced by the choice of demonstration sets. Exhaustive searches over these sets are combinatorial, and existing selectors often rely on relevance or likel...
This work introduces prediction-powered quantile learning: a synthetic-labeled pool estimates pinball risk, paired trusted and synthetic outcomes correct its bias, and an independent trusted split performs final conformalization.
Machine learning task type identification is essential for constructing valid ML pipelines, yet in practice it is typically specified manually. We investigate whether large language models (LLMs) can infer both the data domain and the downstream prediction task directly from dataset-level information when only the targ...
Petros Tsialis, Steffen Limmer, Tobias Rodemann et al.· 0 citations
An expert-driven methodology for constructing such alignment data and applying it to a normative framework grounded in Islamic ethical, theological, and jurisprudential traditions is presented.
H. Sencar, Rezart Beka, D. Naeem et al.· 0 citations
Large language models (LLMs) can sometimes report perturbations to their internal activations---even when the input provides no evidence that an intervention occurred. How do models detect and localize such internal changes? We study this question using a controlled task that keeps the input text fixed. We either injec...
Jia-Hong Zou, Xiang-Kun Sun, Ling-Kai Kong et al.· 0 citations
Bipartite matching is a fundamental problem in game theory and market design. Classical approaches such as Gale--Shapley assume complete preferences and centralized computation, whereas many real-world matching processes are decentralized, asynchronous, and shaped by sequential interaction under limited information. We...
Wang-Xuan Fan, Xiao-Yu Nie, Zhou-Tian Shi et al.· 0 citations
Codoku (code sudoku), a renewable benchmark in which a solver fills typed cells in a partial program to satisfy global static and dynamic constraints, such as a prescribed control-flow graph and execution path, is introduced.
This work proposes ABC-Align, leveraging abundant pseudo label signal to minimize variance and applying a lightweight, adaptive correction grounded in the human-labeled subset, and empirically demonstrates that ABC-Align achieves superior performance over prior semi-supervised baselines in a series of experiments on an...
Eric Frankel, Bang-Hua Zhu, Sewoong Oh et al.· 0 citations
