Menu
Skills dans ce dépôt
SkillsMP a collecté 1 228 skills depuis ADu2021/skillXiv. Ouvrez un skill pour examiner sa source et ses détails.
ADu2021/skillXivAffichage de 40 skills collectés sur 1 228.
Train text-to-image diffusion models directly in frozen DINOv3 feature spaces, eliminating VAE-based compression. Enables high-resolution synthesis by leveraging VFM representations as native latent manifolds with unified cross-modal transformers.
Langue du texte source : anglais
Generate problem variants from correct model solutions while preserving answer equivalence, enabling self-play training that maintains output diversity and prevents entropy collapse.
Langue du texte source : anglais
Train LLM-based agents for multi-turn SWE tasks via rejection fine-tuning and DAPO RL, scaling to 131k context length achieving 39% Pass@1.
Langue du texte source : anglais
Framework orchestrating competitive debate among specialized agents with different reasoning perspectives. Generates multiple fault propagation traces via code dependency graphs, then resolves to consolidated fixes through structured multi-round competition.
Langue du texte source : anglais
Framework that distills reusable experience from prior agent trajectories enabling continuous learning across issues. Achieves 73% resolution on SWE-Bench by leveraging multi-level experience banks capturing both successful and failed repair attempts.
Langue du texte source : anglais
Automate construction of GitHub issue resolution benchmarks via multi-agent LLM coordination, binary file recovery, and exit-code-based validation, achieving 337 valid instances at $0.047 per instance.
Langue du texte source : anglais
Achieve state-of-the-art software issue resolution through lightweight supervised fine-tuning with structured datasets and curriculum learning, plus test-time scaling—reaching 52.6% on SWE-Bench Verified and 58.8% with test-time strategies, outperforming…
Langue du texte source : anglais
Implement techniques from SWE-Pruner: Self-Adaptive Context Pruning for Coding Agents. LLM agents have demonstrated remarkable capabilities in software development, but their performance is hampered by long interaction contexts, which incur high API costs and…
Langue du texte source : anglais
Improve SWE agent RL via execution-free reward models optimized for three metrics beyond TTS: discriminative ability (AUC) and calibration (ECE). Shows TTS alone insufficient—models must distinguish correct/incorrect trajectories and align confidence with…
Langue du texte source : anglais
Automatically generate executable software engineering environments from GitHub pull requests at million-scale using an autonomous building agent. Detects superficial verification patterns to force genuine code execution testing.
Langue du texte source : anglais
Replace Docker environments with learned LLM surrogates comprising a lightweight sandbox for deterministic file operations, a transition model predicting execution feedback, and a reward model acting as virtual test runner. Reduces training infrastructure…
Langue du texte source : anglais
Enable multimodal models to dynamically switch between text and vision reasoning modes, allocating computation based on perceived difficulty and image resolution, achieving strong performance on both vision-dense and text-heavy benchmarks.
Langue du texte source : anglais
Dynamically switch between explicit reasoning and latent computation modes during inference to optimize the trade-off between reasoning quality and token consumption.
Langue du texte source : anglais
Enable VLMs and LLMs to self-improve at world modeling by treating actions as latent variables and reciprocally optimizing forward and inverse dynamics models using only unlabeled state transitions.
Langue du texte source : anglais
Adapt web agents to new domains through targeted synthetic data generation and quality-aware refinement—identifying and correcting hallucinations while preserving task consistency to enable efficient adaptation with minimal human supervision.
Langue du texte source : anglais
Scale visual reasoning via automated synthesis of challenging questions from seed samples, using verification mechanisms to ensure correctness and verify RL training gains on out-of-domain visual math tasks.
Langue du texte source : anglais
Personalize diffusion models to learn a concept from a single image without overfitting by using timestep-dependent rank masking and orthogonal weight initialization, enabling faithful concept reproduction while maintaining text-guided control.
Langue du texte source : anglais
Accelerate diffusion language models from many to few generation steps using trajectory self-distillation. Collect clean-noisy state pairs along teacher trajectories, apply reverse-KL mode-seeking distillation, and weight losses by decoding order to…
Langue du texte source : anglais
Improve language model performance on complex numerical reasoning over tables through three-stage pipeline: decompose questions into sub-questions, sanitize and clean table data, then generate executable Python code for precise computation.
Langue du texte source : anglais
Improve LLM agents operating under strict feasibility constraints (budget limits, tool usage caps) by separating planning from execution. Generate multiple candidate plans, merge into plan graph, then use external solver (ILP) to find optimal feasible path.…
Langue du texte source : anglais
Design capability-adaptive curricula for code generation by constructing per-problem test suites (basic, intermediate, complex, edge cases) and dynamically weighting training signals based on model capability rather than problem difficulty alone. Optimal…
Langue du texte source : anglais
Token-adaptive preference optimization framework using min-max formulation to reduce multimodal LLM hallucination. Achieves 50% hallucination reduction using min-max distributional robustness while preserving visual grounding.
Langue du texte source : anglais
Build process reward models for tabular reasoning by grounding intermediate reasoning steps in executable tool operations. Train on 60K+ step-level annotations combining verification rationales with tool execution feedback. TaTToo achieves 30.9% improvement…
Langue du texte source : anglais
Route queries to multiple specialized agents dynamically using reasoning-aware routing that generates natural-language justification before predicting candidate agents. Enables seamless addition of new agents without system redesign. Routes aggregate…
Langue du texte source : anglais
Leverages temporal dynamics in diffusion models by aggregating predictions across denoising steps for improved inference quality without retraining.
Langue du texte source : anglais
Adapt pretrained video diffusion models to conditional tasks using only 10-30 samples without architectural changes.
Langue du texte source : anglais
Improves self-rewarding language models through temporal framework that prevents representational collapse by anchoring rejected responses and guiding chosen responses across training phases.
Langue du texte source : anglais
Evaluate agents on 89 challenging terminal-based tasks where frontier models score below 65%, providing realistic assessment of command-line interface automation capability. Use when benchmarking agents designed for system administration, automation, or…
Langue du texte source : anglais
Learn when to terminate chain-of-thought reasoning by predicting answer arrival patterns. Train an early-exit mechanism on empirical first-answer positions to reduce reasoning length by 14-55% while maintaining or improving accuracy.
Langue du texte source : anglais
Apply tabu search enhanced with short-term tabu lists and long-term elite memory for simulation optimization with expensive, noisy evaluations. Balances exploration (avoiding cycling) and exploitation (leveraging best solutions). Use for multimodal landscapes…
Langue du texte source : anglais
Applies test-driven development to agent prompts by iteratively refining prompts against behavioral test suites until compliance is achieved. Enables measurable agent behavior validation through semantic mutation testing and specification evolution.
Langue du texte source : anglais
Adapt vision-language models at inference without labeled data by generating multiple predictions and rewarding high-frequency outputs. Trigger: improve VLM accuracy on deployment with self-generated supervision signals.
Langue du texte source : anglais
Scale model performance at test time by generating multiple reasoning trajectories and selecting the best using a self-supervised process reward model. MetaStone-S1 achieves 32B-equivalent performance using only 32B parameters and 53M for trajectory scoring,…
Langue du texte source : anglais
Generate high-quality videos through iterative test-time refinement: agents critique outputs on visual/audio/context fidelity and a reasoning agent synthesizes feedback to improve prompts across multiple generations.
Langue du texte source : anglais
Enables agents to synthesize, verify, and evolve executable tools during inference rather than relying on static tool libraries, improving reasoning on heterogeneous scientific domains through dynamic tool adaptation.
Langue du texte source : anglais
Restore degraded images while preserving textual fidelity using TeReDiff, a multi-task diffusion framework integrating text spotting with U-Net features and VLM-verified dataset curation.
Langue du texte source : anglais
Modern CI/CD pipelines integrating agent-generated code exhibit a structural failure in responsibility attribution. Decisions are executed through formally correct approval processes, yet no entity possesses both the authority to approve those decisions and…
Langue du texte source : anglais
Systematic evaluation toolkit for assessing large language models across multiple dimensions, enabling comprehensive benchmarking of agent capabilities and comparative analysis of model performance.
Langue du texte source : anglais
Reduce reasoning model inference cost by selectively applying latent iterations—use a lightweight neural decider to identify hard tokens and refine only those with focused LoRA modules, achieving 94% token exemption with 8-11% accuracy gain.
Langue du texte source : anglais
Dynamically allocate reasoning budgets per-task: 36.8% length reduction while improving accuracy 8.4% by learning which problems warrant deep reasoning vs. quick answers. Use when optimizing reasoning efficiency across variable-difficulty problems.
Langue du texte source : anglais