메뉴
이 저장소의 skills
SkillsMP는 ADu2021/skillXiv에서 1,228개의 skill을 수집했습니다. skill을 열어 소스와 세부 정보를 확인하세요.
ADu2021/skillXiv수집된 skill 1,228개 중 40개를 표시합니다.
Compact Mentalese symbolic format trained via SFT, then refined with SLPO (Shorter Length Preference Optimization) to reward concise correct solutions without over-penalizing necessarily long reasoning. Compress reasoning while maintaining accuracy.
원문 언어: 영어
Detect safety violations in mobile GUI agents using dual mechanisms: formal verification rules for system-level risks (privilege escalation, credential exposure) and VLM-based contextual judges evaluating action appropriateness, achieving 10-30% safety…
원문 언어: 영어
Enable reasoning models to adaptively switch between fast direct responses and slow detailed reasoning, reducing token consumption by 15-40% while maintaining accuracy through dual-mode fine-tuning.
원문 언어: 영어
Diagnose and prevent over-searching—unnecessary search invocations that degrade LLM performance and waste tokens. Framework introduces Tokens Per Correctness (TPC) metric and OverSearchQA dataset to train models that know when NOT to search. Teaches agents to…
원문 언어: 영어
Extract and manipulate internal progress vectors in reasoning models to accelerate thinking phases while maintaining answer quality, achieving 80%+ improvements in token efficiency.
원문 언어: 영어
A 3B unified model combining image understanding, text-to-image generation, and image editing end-to-end rather than as separate frozen components. Use when you need a single efficient model for multiple vision-language tasks without the overhead of separate…
원문 언어: 영어
Weight distillation loss by student pass rate using Beta kernel: suppress mastered (high pass rate) and intractable (low pass rate) problems, prioritize intermediate zone where learning signal is richest.
원문 언어: 영어
Improves LLM-based evolutionary search by addressing context pollution, mode collapse, and weak collaboration through hierarchical context management, momentum-based backtracking, and adaptive sampling policies.
원문 언어: 영어
PaCoRe: Learning to Scale Test-Time Compute with Parallel Coordinated Reasoning. From arXiv:2601.05593
원문 언어: 영어
Verify solution quality through pairwise comparison rather than pointwise scoring. Implement topology coverage and Swiss refinement to allocate verification compute to uncertain pairs, improving calibration and reducing verification overhead.
원문 언어: 영어
Writing effective rebuttals is a high-stakes task that demands more than linguistic fluency, as it requires precise alignment between reviewer intent and manuscript details. Current solutions typically treat this as a direct-to-text generation problem,…
원문 언어: 영어
Advanced reasoning approach for optimizing inference efficiency through meta-cognitive planning, enabling agents to make better decisions with reduced computational overhead.
원문 언어: 영어
Monitor parallel reasoning branches via 2D probing—periodically extracting intermediate answers to detect consensus and prune divergent branches. Reduces token cost by 25%+ while maintaining accuracy without requiring model retraining.
원문 언어: 영어
Train language models to explore multiple reasoning paths simultaneously via reinforcement learning. Uses progressive curriculum learning to address cold-start problems, enabling 8.4% accuracy gains over sequential reasoning on complex mathematical tasks.
원문 언어: 영어
Linear RNN architecture enabling efficient parallel computation over directed acyclic graphs with stabilization modes for long-range dependencies without sequential traversal.
원문 언어: 영어
Generate 3D objects with explicit part structures enabling compositional editing and animation. Decouples structure planning from geometry synthesis using two stages: autoregressive bounding box generation and part-aware refinement.
원문 언어: 영어
Generates semantically-meaningful 3D parts from single images via compositional diffusion transformers with part-level identity and local-global attention.
원문 언어: 영어
Use Pass@k as the reward metric in RL training to balance exploration and exploitation, enabling models to learn diverse sampling strategies while maintaining correctness.
원문 언어: 영어
Improve post-RL performance by reweighting SFT loss using importance sampling. Prioritize training examples that match the target policy distribution, not the behavior policy. Achieves 14.6% Pass@8 gains on AIME.
원문 언어: 영어
Systematic evaluation toolkit for assessing large language models across multiple dimensions, enabling comprehensive benchmarking of agent capabilities and comparative analysis of model performance.
원문 언어: 영어
Comprehensively evaluate 12+ parameter-efficient fine-tuning methods for RL with Verifiable Rewards (RLVR). Show DoRA/AdaLoRA outperform LoRA, SVD-based methods fail on RL, extreme reduction creates bottlenecks—providing empirical evidence that…
원문 언어: 영어
Enable language models to actively manage their context using memory tools (read, index, note-take, delete). Models receive agency to self-engineer context rather than passively consume pre-staged information, maintaining compact high-quality reasoning states…
원문 언어: 영어
Optimize multimodal LLMs by directly targeting perception errors using KL-divergence based perception loss, improving visual reasoning by 8-19% on vision-dependent tasks. Integrates perception-aware signals into policy gradients without relying solely on…
원문 언어: 영어
Enable language models to reason over extremely long contexts (128K tokens) by encoding context into lightweight LoRA adapters during test time, achieving 20% performance improvements without full model retraining.
원문 언어: 영어
Method to identify and manipulate interpretable directions in model activation space corresponding to personality traits. Enables real-time monitoring of trait drift and preventive steering to mitigate unwanted behavioral shifts in language models.
원문 언어: 영어
Optimize sampling budget allocation for self-consistency inference by treating trajectory allocation as a crowdsourcing problem. Introduce self-consistency rate (agreement with infinite-budget consensus) as optimization target. Offline setting uses Optimistic…
원문 언어: 영어
Physical principles are fundamental to realistic visual simulation, but remain a significant oversight in transformer-based video generation. This gap highlights a critical limitation in rendering rigid body motion, a core tenet of classical mechanics. While…
원문 언어: 영어
Apply reinforcement learning to flow-based VLA models by modeling denoising as an MDP and converting ODEs to SDEs, enabling efficient policy optimization for robotic control without expensive supervised fine-tuning data collection.
원문 언어: 영어
Accelerate diffusion transformers through training-free sparse attention combining exact computation for critical blocks with efficient approximation for non-critical ones. Achieves 2-3× speedup without retraining.
원문 언어: 영어
Enhance Masked Autoencoders through three algorithmic improvements: deeper decoders enabling semantic-focused encoding, larger masking blocks providing richer context, and multiple class tokens capturing diverse global properties. Combine with 2B web-crawled…
원문 언어: 영어
Decompose complex queries into sub-questions and allocate computational budgets adaptively based on estimated difficulty, achieving 70% accuracy improvements and 39% token reduction without retraining.
원문 언어: 영어
Develop reasoning capabilities in multimodal models through interactive gameplay rather than supervised learning, achieving superior transfer to diverse reasoning tasks.
원문 언어: 영어
Generate spatially and temporally coherent multi-view video through autoregressive conditioning with camera-guided retrieval and progressive context scaling. Enable long-video generation maintaining spatio-temporal memory across viewpoint changes.
원문 언어: 영어
Improve LLM training stability and generalization by reparameterizing weight matrices as orthogonal transformations, achieving better perplexity than AdamW with fewer trainable parameters.
원문 언어: 영어
Use language models as optimizers to iteratively improve complex systems (prompts, code, agent policies) through noisy feedback and sampling. POLCA maintains a priority queue of candidates, explores with epsilon-nets, and learns meta-insights from trial…
원문 언어: 영어
Learn generalizable reward models via unsupervised policy discrimination: pretraining models to distinguish between different policies enables efficient adaptation to human preferences and strong RLHF performance.
원문 언어: 영어
Remove deep layers during context encoding (prefill) while keeping them for token generation (decode). Identifies layer importance asymmetry via virtual gates; achieves 1.37x prefill speedup without retraining on any pre-trained model.
원문 언어: 영어
Improve speculative decoding throughput by employing position-specialized draft layers that handle position-specific error accumulation patterns.
원문 언어: 영어
Train LLM agents via multi-turn reinforcement learning by systematically optimizing environment complexity, reward signals, and policy initialization. Use curriculum learning, dense verified rewards, and domain-specific SFT for reliable agent convergence…
원문 언어: 영어
Optimize text-to-image models using pairwise preference comparisons instead of pointwise rewards, eliminating normalization instability and reward hacking while providing fine-grained image quality differentiation
원문 언어: 영어