Menu
Skills in this repository
SkillsMP has collected 1,228 skills from ADu2021/skillXiv. Open a skill to review its source and details.
ADu2021/skillXivShowing 40 of 1,228 collected skills.
Accelerate video generation and enable long-video synthesis by decomposing into two diffusion stages: first generate compact semantic features for global planning, then generate VAE latents conditioned on semantics. Includes learnable semantic compression to…
Encode reasoning steps as hidden embeddings instead of explicit text using contrastively-trained sentence transformers and lightweight distilled models, reducing token generation cost while preserving semantic alignment with ground-truth reasoning.
Build vision-language agents that seamlessly integrate visual reasoning with dynamic tool manipulation (search, cropping) through reinforcement learning, achieving state-of-the-art performance on fine-grained visual understanding tasks—surpassing proprietary…
Deploy a state-of-the-art binary classifier using ModernBERT to detect prompt injection attacks and protect LLMs from adversarial input manipulation.
Implement adaptive parallel decoding for language models using diffusion-based next-sequence prediction. Enable dynamic block-based token generation with confidence thresholds to achieve 2x+ speedups while maintaining competitive performance. Retrofit…
Replace parallel self-consistency with sequential reasoning where chains iteratively build on previous attempts, weighted by inverse entropy to prioritize confident solutions, achieving 46.7 pp accuracy gains over parallel approaches.
Accelerate language model generation 3-5x by combining autoregressive and masked token prediction. Works via fine-tuning—no architectural changes needed. Parallel decode non-consecutive tokens with entropy-bounded sampling.
Demonstrate that synthetic CoT traces with incorrect final answers outperform human-written correct solutions for supervised fine-tuning. Distribution proximity between training data and student model's natural output matters more than correctness—validating…
DeltaLoss sensitivity metric combining gradient and quantization-induced parameter deviation for adaptive bit-width allocation, with lightweight pre-tuning search for scale initialization, enabling competitive accuracy at 4-5 bits in 2.5-6 hours.
Convert videos to language-based representations and leverage LLM reasoning without video-specific training.
Train LLMs to reason implicitly with step-level supervision, stabilizing latent representations while preserving 2.3× inference speedup over explicit chain-of-thought. Addresses training collapse in implicit reasoning by aligning intermediate latent states…
Improve pass@K by using asymmetric probability boosting: increase probabilities of top-K correct solutions while penalizing top-1 incorrect predictions. Focus boosting on high-entropy tokens where exploration helps most.
Place RMSNorm immediately after every linear layer to stabilize activation scales at O(sqrt(d)) and reduce Hessian spectral norm. Enables 3-10x larger learning rates and faster convergence without architectural changes; improves loss by 0.08 on 7B models.
Implement efficient memory systems for long-term LLM agent interactions using semantic compression, achieving 30-fold inference token reduction while improving F1 scores by 26.4%—enabling agents to learn from extended interaction histories without prohibitive…
Train LLMs for multi-turn tool-integrated reasoning end-to-end using RL without supervised pretraining. SimpleTIR stabilizes training by filtering void turns (responses lacking code blocks or final answers) to prevent gradient explosion from distributional…
Apply reinforcement learning to Vision-Language-Action models for robotic control, achieving 99% LIBERO task success and discovering novel manipulation strategies (pushcut) without task-specific reward engineering. Scales efficiently via parallelized…
Evaluating whether multimodal large language models truly understand long-form scientific papers remains challenging: answer-only metrics and synthetic 'Needle-In-A-Haystack' tests often reward answer matching without requiring a causal, evidence-linked…
Replace LoRA's two-matrix decomposition with a single learnable matrix (AA⊤) to eliminate scale imbalances and improve training stability. Reduces parameters by ~50% while maintaining or exceeding LoRA performance.
Improve agent performance by autonomously distilling behavioral patterns from trajectories into reusable skills, then using these skills to guide future decisions. Achieves 89.9% success on ALFWorld through differential processing of success vs failure…
Adapt full-attention language models to sliding window attention without expensive retraining. Combine five synergistic strategies (full decode, sink tokens, interleaved layers, chain-of-thought, fine-tuning) achieving 30-100% speedups while maintaining…
Optimize model preferences by decoupling preference learning from generation quality. Explicitly maximize chosen response likelihood while using token-level stabilization to prevent quality degradation from over-suppressing rejected responses.
Design heterogeneous agentic systems combining specialized small models with selective large model deployment for superior economics and performance.
Stabilize RL for LLM reasoning via three-phase decomposition: fast inner trajectory optimization, repositioning to manage off-policy drift, slow correction for stable updates. Achieve up to 2.80-point math reasoning gains over GRPO while reducing rollouts…
Improve LLM search agents by optimizing query quality at each step using process-level rewards. Framework teaches agents to iteratively refine search queries through imitation, alignment, and generalization stages. Agents learn to identify low-quality queries…
Shift agent verification from post-hoc external judgment to proactive in-situ self-evidence curation. Agents generate atomic evidence tuples during execution, guided by 3C principles (Completeness, Conciseness, Creativity), with structured verifier feedback…
Deploy compact vision-language-action models that run on consumer GPUs for natural language robot control.
Replaces iterative diffusion with single-step decoding for image compression. Combines VAE latents with fidelity guidance and rate annealing training. Achieves 20× decoding speedup with improved perceptual quality.
Enable policy gradient optimization on soft LLM tokens by injecting Gumbel noise and applying Gumbel-Softmax reparameterization—allowing soft-thinking patterns to match discrete-token RL performance while maintaining continuous optimization advantages.
Defends tool-augmented LLM agents against prompt injection via iterative input sanitization. Multi-pass inspection detects malicious instructions in untrusted data, remediates them, and re-evaluates until clean or iteration limit reached. Raises attack…
Optimize MoE training through memory-efficient backward pass, IO-aware kernel design overlapping memory operations, and token rounding routing. Avoid caching large-scale activations, fuse operations with GEMM, implement ping-pong scheduling. Achieve 1.86×…
Enhanced language model pre-training methodology improving linguistic competence across languages, strengthening foundational capabilities for multilingual agent systems.
Train socially intelligent LLMs via utterance-level credit assignment and multi-dimensional reward aggregation for social interactions.
Improve vision-language model reasoning efficiency by decoupling perception (identifying task-relevant image regions) from reasoning (generating explanations), enabling asymmetric compute allocation. Reduces token overhead while improving accuracy through…
Train LLMs to simultaneously act as reasoning agents and reward models through recycled on-policy rollouts, eliminating separate reward infrastructure while achieving 9.7% gains on reasoning tasks.
Train process reward models without ground-truth references using synthetic verification data from generators and verifiers. SPARK achieves 67.5 F1 on ProcessBench—ideal when step-level annotations are expensive but verification is available.
Learn sparse attention patterns for reasoning model decoding via self-distilled gating, achieving 9x speedup at 90% sparsity while maintaining reasoning quality.
Accelerate masked discrete diffusion models by dynamically truncating unnecessary masked tokens while maintaining complete representation through positional information. Use register tokens as compressed representations, implement step-causal attention masks…
Deploy frontier-level reasoning with only 11B active parameters using sparse MoE with 288 routed experts plus shared expert. Use Metropolis Independence Sampling-Filtered Policy Optimization (MIS-PO) to stabilize RL training at scale, replacing continuous…
Overcome the attention gap in sparse transformers by training with both full and sparse attention simultaneously, aligned through bidirectional losses that encourage naturally sparser distributions while maintaining learning capability, enabling efficient…
Accelerate video diffusion transformer inference by 1.58-1.85× through discovering and exploiting sparse attention patterns that exhibit diagonal, multi-diagonal, and vertical-stripe structures.