← all publishers

adu2021

@adu2021 source repo

Report this profile Claim this profile

• 1.2k skills
• 0 followers
• 1 day ago last updated

1. ▌
   Heteroscale Autoscaling · adu2021

   Scale disaggregated LLM inference (prefill-decode) via topology-aware scheduling and metric-driven policies, achieving 26.6% GPU utilization improvement and conserving hundreds of thousands GPU-hours daily
   0
   installs
2. ▌
   Hybrid Linear Attention · adu2021

   Design hybrid architectures mixing linear and full attention at optimal ratios. Superior standalone linear models don't necessarily excel in hybrids; recall improves significantly with more full attention layers at ratios below 3:1, enabling efficient long-sequence processing.
   0
   installs
3. ▌
   Latent Chain Of Thought · adu2021

   Improve reasoning quality by decoupling reasoning from verbalization. Planner generates deterministic latent trajectories while Decoder grounds them to text. Enables dynamic termination, better Pass@k scaling, and interpretable intermediate states.
   0
   installs
4. ▌
   Learning Beyond Teacher · adu2021

   Train student models to exceed teacher performance by extrapolating rewards during distillation using a scaling factor λ1. Enables unified students to surpass all individual domain teachers through controlled reward amplification.
   0
   installs
5. ▌
   Let Late Early Training · adu2021

   Accelerate LLM pretraining by leveraging small pretrained models as teachers, enabling large models to explicitly learn late-layer knowledge in early layers via alignment loss, achieving 1.6x convergence speedup and 5% downstream improvement even with 10x smaller teachers.
   0
   installs
6. ▌
   Math Reasoning Transfer · adu2021

   Understand why math reasoning improvements don't always transfer to general capabilities. Use RL-based training instead of SFT to preserve representation structure and enable broader generalization.
   0
   installs
7. ▌
   Memory Agent Evaluation · adu2021

   Evaluate and improve memory capabilities in LLM agents across four competencies: accurate retrieval, test-time learning, long-range understanding, and selective forgetting. Identifies critical gaps in how agents store, update, and revise information.
   0
   installs
8. ▌
   Memory Operating System · adu2021

   Treat memory as a manageable system resource for LLMs through unified management of plaintext, activation, and parameter-level memories with dynamic scheduling and lifecycle governance.
   0
   installs
9. ▌
   Mimo Vl Vision Language · adu2021

   Build state-of-the-art 7B multimodal models by combining four-stage vision-language pretraining with mixed on-policy RL integrating verifiable and human feedback rewards.
   0
   installs
10. ▌
    Minicpm4 Efficient Llms · adu2021

    Build ultra-efficient language models for edge devices using sparse attention, high-quality data filtering, and ternary quantization, achieving Qwen3-8B performance with 22% of training tokens.
    0
    installs
11. ▌
    Nestbrowse Info Seeking · adu2021

    Scale information-seeking agents via nested architecture separating outer reasoning from inner page exploration. Minimal toolkit (search, visit, click, fill) handles dynamic web interactions, inner loop filters content before context injection, multi-task learning trains outer/inner jointly—enabling 4B models to match larger competitors.
    0
    installs
12. ▌
    Neuralos Gui Simulation · adu2021

    Simulate GUI behavior by predicting screen frames in response to user inputs. NeuralOS combines hierarchical RNNs for state tracking with diffusion-based rendering, capturing mouse interactions and application state transitions. Trains on synthetic demonstrations plus random exploration; achieves 50-61% human indistinguishability on basic operations while maintaining 18 fps inference on single H100.
    0
    installs
13. ▌
    Octopus Self Correction · adu2021

    Enable vision-language models to self-correct by synthesizing dense training examples from existing rollouts, creating n² correction pairs from n original trajectories.
    0
    installs
14. ▌
    Omni Moe Atomic Experts · adu2021

    Scale mixture-of-experts models efficiently by decomposing experts into atomic vector pairs with Cartesian product routing and expert-centric scheduling. Achieves 10.9× speedup and 50% fewer parameters versus fine-grained baselines through system-algorithm codesign that converts scattered memory access into contiguous batched operations.
    0
    installs
15. ▌
    Pref Grpo Text To Image · adu2021

    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
    0
    installs
16. ▌
    Prism Diffusion Scaling · adu2021

    Scale inference efficiency for discrete diffusion language models through hierarchical trajectory search with adaptive pruning and self-verified feedback. Achieve 3-4× speedup versus best-of-N with equal quality.
    0
    installs
17. ▌
    Recode Unified Planning · adu2021

    Unifies planning and action by treating plans as abstract placeholder functions recursively decomposed to primitive actions. Enables agents to dynamically adjust abstraction levels per task without rigid hierarchies. Improves inference performance and training efficiency through automatic multi-level data generation.
    0
    installs
18. ▌
    Relic Video World Model · adu2021

    Compressed historical latents with camera poses in KV cache (4× compression), extended teacher training (20-second sequences), and replayed back-propagation (block-wise differentiation) enabling real-time interactive video generation with long-range spatial consistency.
    0
    installs
19. ▌
    Sealqa Reasoning Search · adu2021

    Evaluate search-augmented language models on fact-seeking questions with conflicting or unhelpful search results, revealing critical reasoning gaps in frontier models and testing robustness to noisy information.
    0
    installs
20. ▌
    Search R2 Refinement Rl · adu2021

    Decompose search-integrated reasoning into actor (generates trajectories) and refiner (identifies and corrects flawed steps). Refiner performs surgical cut-and-regenerate correction preserving valid prefixes. Trained end-to-end with GRPO; adds minimal overhead (2%) to baseline.
    0
    installs
21. ▌
    Sparse Lavida Diffusion · adu2021

    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 for bidirectional context. Achieve 1.95-2.83× speedups across generation tasks.
    0
    installs
22. ▌
    Sparse Sparse Attention · adu2021

    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 inference without capability degradation.
    0
    installs
23. ▌
    Sparser Block Attention · adu2021

    Accelerates self-attention by reordering tokens to concentrate relevant tokens within fewer blocks. Achieves 2.75x speedup on long-context prefilling by permuting tokens so semantically related information clusters together, enabling aggressive block sparsity without accuracy loss.
    0
    installs
24. ▌
    Tower Plus Multilingual · adu2021

    Build multilingual LLMs that excel at machine translation while preserving general-purpose capabilities through a staged training pipeline combining continued pretraining, SFT, preference optimization, and RL with verifiable rewards.
    0
    installs
25. ▌
    Wan Move Motion Control · adu2021

    Enable precise motion control in video generation using dense point trajectories as latent space features. Wan-Move integrates with existing models without architecture changes—ideal when you need fine-grained scene control without auxiliary motion encoders.
    0
    installs
26. ▌
    Weight Decay Plasticity · adu2021

    Improve downstream task performance by increasing weight decay during pretraining (0.3-1.0 vs. default 0.1), enabling better model plasticity and representation structure.
    0
    installs
27. ▌
    Manifold Aware Rl Video · adu2021

    Constrain video GRPO policy updates to stay within pre-trained model's data manifold using dual-control exploration. Implement precise manifold-aware SDE with logarithmic noise variance correction (captures geometric signal decay standard methods miss). Apply gradient norm equalizer to balance learning across diffusion timesteps (mitigate vanishing/exploding gradients). Use dual trust region combining position control (anchored exploration) and velocity control (KL constraints) for stability-plasticity balance.
    0
    installs
28. ▌
    Agent Workplace Learning · adu2021

    Implements The Agent's First Day from arXiv:2601.08173
    0
    installs
29. ▌
    Agent0 Vl Self Evolution · adu2021

    Enable vision-language agents to self-evolve by grounding verification in tool outputs rather than text: implement nested loops where Solver+Verifier generate trajectories and tool-based feedback, then optimize via GRPO using self-generated rewards without external supervision.
    0
    installs
30. ▌
    Agentic LLM Data Science · adu2021

    Train agentic LLMs through curriculum-based learning to autonomously execute full data science workflows from raw data to analysis reports, enabling 8B models to match proprietary systems.
    0
    installs
31. ▌
    Art Of Scaling Test Time · adu2021

    Comprehensive empirical study recommending model-specific test-time scaling strategies (majority voting, first-finish search) across eight LLMs based on architectural family, problem difficulty, and compute budget rather than universal approaches.
    0
    installs
32. ▌
    Auto Codebench Generator · adu2021

    Automatically generates diverse multilingual code benchmarks using LLMs, creating 3920 problems across 20 programming languages with quality assurance filtering.
    0
    installs
33. ▌
    Bavt Budget Aware Search · adu2021

    Allocate LLM reasoning budget optimally via value tree search: use residual value prediction to estimate step utility, then dynamically shift exploration-exploitation balance as budget depletes. Outperform high-budget baselines at 1/4 cost.
    0
    installs
34. ▌
    Clift Light Field Tokens · adu2021

    Represent 3D scenes as compressed light-field tokens for efficient neural rendering. Multi-view images are tokenized via Plücker coordinates, condensed through K-means clustering, and rendered adaptively. Achieves 5-7× data reduction versus MVSplat while enabling on-the-fly quality-speed tradeoffs: up to 66% FPS improvement with controlled token counts.
    0
    installs
35. ▌
    Code2world Gui Synthesis · adu2021

    Predict GUI state evolution by generating HTML code rather than pixel images. Combines visual fidelity of pixel-based approaches with structural precision of code-based methods through deterministic rendering. Enables agents to evaluate action consequences and select best decisions before execution.
    0
    installs
36. ▌
    Confidence Rl Finetuning · adu2021

    Improve language model reasoning using only model confidence as reward signals, eliminating need for labels or preference models while achieving substantial gains with minimal data.
    0
    installs
37. ▌
    Crisp Concept Unlearning · adu2021

    Permanently remove unwanted concepts from LLMs by identifying and suppressing sparse autoencoder features across layers, creating parameter-level changes that prevent reversal.
    0
    installs
38. ▌
    Dino R1 Vision Reasoning · adu2021

    Apply reasoning-focused RL to vision foundation models using GRPO to develop deep visual understanding and abstract reasoning beyond visual recognition.
    0
    installs
39. ▌
    Docdancer Document Agent · adu2021

    Build open-source agents for document question-answering by modeling DocQA as information-seeking with explicit tool utilization. DocDancer uses an exploration-then-synthesis pipeline to generate high-quality training data, addressing the scarcity that limits agent-based document understanding systems.
    0
    installs
40. ▌
    Expert Threshold Routing · adu2021

    Improve MoE language model efficiency with causal threshold-based routing that eliminates auxiliary losses and enables dynamic per-token computation.
    0
    installs
41. ▌
    Focused Chain Of Thought · adu2021

    Training-free prompting strategy that pre-organizes query information into compact structured format, reducing generation tokens by 2-3× while maintaining reasoning accuracy. Apply when reasoning performance is bottlenecked by verbose input formatting.
    0
    installs
42. ▌
    Glm Multimodal Reasoning · adu2021

    Train vision-language models with curriculum-based reinforcement learning (RLCS) to improve reasoning across diverse multimodal tasks. Dynamically adjust training difficulty to model capability, preventing both trivial and overly-hard examples.
    0
    installs
43. ▌
    Infialign Data Selection · adu2021

    Combines SFT and DPO with robust data selection pipeline using multidimensional quality metrics. Achieves DeepSeek-R1 performance with 12% training data, enabling efficient reasoning model alignment.
    0
    installs
44. ▌
    Jacobi Parallel Decoding · adu2021

    Enable efficient parallel decoding while preserving causal inference properties through progressive distillation with cyclic noise scheduling. Reduce training complexity via block-wise sparse attention, implement rejection recycling and multi-block decoding. Achieve 3.8-4.0× wall-clock speedup on coding benchmarks.
    0
    installs
45. ▌
    Janus Moe Disaggregation · adu2021

    Enable scalable MoE inference by disaggregating attention and expert layers onto independent GPU sub-clusters. Use adaptive two-phase communication, activation load-balanced scheduling, and activation-aware expert management. Achieve 3.9× higher per-GPU throughput than state-of-the-art systems.
    0
    installs
46. ▌
    Llada2 Diffusion Scaling · adu2021

    Convert pre-trained autoregressive models into large-scale diffusion language models via Warmup-Stable-Decay training strategy. Progressively increase block size during warmup, perform stable diffusion training, then decay to smaller blocks for inference. Achieve 535 tokens-per-second with document-level masking and confidence-aware training.
    0
    installs
47. ▌
    Medical Video Generation · adu2021

    Generate accurate, high-quality medical videos for clinical education and documentation by leveraging large-scale annotated medical datasets with domain-specific fine-tuning on video diffusion models.
    0
    installs
48. ▌
    Multiagent Communication · adu2021

    Enable agents to communicate through shared latent thoughts rather than natural language, recovering both shared and private latent representations with theoretical guarantees for more efficient collaboration.
    0
    installs
49. ▌
    Native Parallel Reasoner · adu2021

    Enable LLMs to develop genuine parallel reasoning without external supervision through progressive self-distilled training. Transform models from sequential reasoning to native parallel cognition with 4.6× speedup—ideal when latency and reasoning quality both matter.
    0
    installs
50. ▌
    Open Cua Computer Agents · adu2021

    Scales computer-use agent capabilities through reflective Chain-of-Thought reasoning in large-scale annotated datasets spanning multiple operating systems and 200+ applications.
    0
    installs
51. ▌
    Part Aware 3d Generation · adu2021

    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.
    0
    installs
52. ▌
    Pixio Masked Autoencoder · adu2021

    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 images and soft self-curation for competitive downstream performance.
    0
    installs
53. ▌
    Pop Prefill Only Pruning · adu2021

    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.
    0
    installs
54. ▌
    Proact Agentic Lookahead · adu2021

    Learn to ground LLM agent planning in real environment dynamics using Monte-Carlo Tree Search exploration combined with lightweight Monte-Carlo critics, reducing hallucination-driven planning failures in interactive tasks.
    0
    installs
55. ▌
    Realmem Memory Benchmark · adu2021

    Implements RealMem from arXiv:2601.06966
    0
    installs
56. ▌
    Reasoning As Compression · adu2021

    Optimize chain-of-thought reasoning under computational budgets using information-theoretic compression principles, improving reasoning efficiency without accuracy loss.
    0
    installs
57. ▌
    Rocket Model Compression · adu2021

    Compress language models by formulating heterogeneous layer-wise allocation as a constrained knapsack problem. Measure per-layer compression error, solve via dynamic programming to minimize total error within target model size, and avoid pathological solutions where some layers are severely damaged.
    0
    installs
58. ▌
    Samtok Mask Tokenization · adu2021

    Tokenize any region mask into two special tokens, enabling pixel-wise tasks like segmentation and region understanding in multimodal language models without architectural changes. Use when adapting vision-language models to perform pixel-level understanding and manipulation.
    0
    installs
59. ▌
    Scaling Latent Reasoning · adu2021

    Scales reasoning depth through internal iteration rather than explicit generation. Ouro models perform repeated computation in latent space with entropy-regularized objectives enabling learned depth allocation. Smaller 1.4B model matches 12B standard models through improved knowledge manipulation.
    0
    installs
60. ▌
    Soft Instruction Defense · adu2021

    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 barrier while maintaining agent usability.
    0
    installs
61. ▌
    Synthrl Visual Reasoning · adu2021

    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.
    0
    installs
62. ▌
    Test Time Tool Evolution · adu2021

    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.
    0
    installs
63. ▌
    Trapo Semi Supervised Rl · adu2021

    Bridge labeled and unlabeled data through trajectory similarity in reinforcement learning. Select reliable unlabeled samples by comparing pass-rate evolution trajectories against labeled data. Achieve 42.6% accuracy with 1K labeled + 3K unlabeled samples, surpassing fully-supervised training on 45K labels with 10% annotation budget.
    0
    installs
64. ▌
    Tree Search LLM Agent Rl · adu2021

    Train LLM agents via tree-search rollout sampling and step-wise advantage estimation. Achieve 1.5x more rollouts within fixed token budgets and implicit step-level preference learning through dual-level advantage computation on tree structures.
    0
    installs
65. ▌
    Trust SQL Rl Text To SQL · adu2021

    Convert natural language to SQL for unknown database schemas by formulating the task as a partially observable MDP. Use dual-track GRPO (token-level masked advantages) to learn schema discovery and query generation jointly.
    0
    installs
66. ▌
    Uniworld Semantic Vision · adu2021

    Combine semantic encoders from multimodal LLMs with contrastive learning to create unified high-resolution encoders for both visual understanding and generation tasks without relying on VAE compression.
    0
    installs
67. ▌
    Video Temporal Reasoning · adu2021

    Diagnose and improve temporal pattern recognition in video-language models using SpookyBench, which isolates temporal information from spatial cues.
    0
    installs
68. ▌
    Dab Data Agent Benchmark · adu2021

    Evaluate data agents on realistic multi-database queries across 54 tasks spanning 12 datasets, 9 domains, and 4 DBMS systems. Reveals frontier models achieve only 38% pass@1 accuracy, with 85% of failures from incorrect planning rather than data selection. Benchmark captures critical challenges: heterogeneous database integration, identifier reconciliation, unstructured text extraction, and domain knowledge. Identifies optimal exploration ratio (~20% of operations) and cost-efficiency differences across models.
    0
    installs
69. ▌
    Memdlm Parametric Memory · adu2021

    Enhance diffusion language model performance on long-context tasks by embedding simulated denoising into training via bi-level optimization. Fast weights capture local trajectory experience; base model optimized with accumulated parametric memory. Achieves +17.0% on RULER Variable Tracking (8K) and +9.6% on BABILong with gains primarily from training-stage improvements.
    0
    installs
70. ▌
    Msft Mixture Overfitting · adu2021

    Identify three ranked findings on multi-task SFT: (1) heterogeneous overfitting—sub-datasets peak at different training points (contradicts uniform duration practice); (2) parameter divergence—excluding 1/10 of data shifts optimal points 0.91 epochs for remaining tasks; (3) SFT compute negligible (0.01% of training). Implement mSFT: iterative roll-out/roll-back search per-dataset. Robust across 0.5B-8B models, 9K-27K samples, 5-15 tasks, achieving +3.4% improvement with reduced FLOPs.
    0
    installs
71. ▌
    Sem Sparse Debiasing Vlm · adu2021

    Achieve post-hoc debiasing of frozen vision-language models by operating in sparse autoencoder latent space. Identify and modulate bias-relevant neurons with neuron-level precision while preserving task-relevant features, requiring no task-specific fine-tuning.
    0
    installs
72. ▌
    Agentic Critical Training · adu2021

    Improves LLM agent decision-making by training agents to first critically evaluate actions before generating, using RL on action-pair comparisons. Develops intrinsic reasoning about action quality without requiring reflection supervision.
    0
    installs
73. ▌
    Alf Load Balancing Theory · adu2021

    Rigorous theoretical framework reformulating DeepSeek's ALF-LB as single-step primal-dual method for assignment problem, proving monotonic Lagrangian improvement, approximate balancing guarantees, and logarithmic expected regret in stochastic settings.
    0
    installs
74. ▌
    Attention Basin Reranking · adu2021

    Demonstrates position bias where LLMs neglect middle content while over-attending to endpoints. Proposes Attention-Driven Reranking (AttnRank) to align content with model's intrinsic attention preferences.
    0
    installs
75. ▌
    Beaver Prompt Compression · adu2021

    Compress long prompts to 1/26th of original size while maintaining retrieval accuracy using hierarchical page-level pooling, without model fine-tuning.
    0
    installs
76. ▌
    Co Evolving Critics Agent · adu2021

    No More Stale Feedback: Co-Evolving Critics for Open-World Agent Learning. From arXiv:2601.06794
    0
    installs
77. ▌
    Coba Rl Budget Allocation · adu2021

    Dynamically allocate training budget across samples using a capability-oriented value function that measures per-sample training importance based on model capability evolution. Reduces training time via greedy heap-based allocation optimizing exploration-exploitation tradeoff.
    0
    installs
78. ▌
    D Core Task Decomposition · adu2021

    Mitigate lazy reasoning in Large Reasoning Models via self-distillation teaching task decomposition, followed by Diversity-Aware GRPO with entropy-based advantage functions, enabling effective decomposition without external teachers while balancing structured reasoning with diversity.
    0
    installs
79. ▌
    Diffucoder Diffusion Code · adu2021

    Train masked diffusion models for code generation using coupled-GRPO to optimize non-autoregressive generation. Achieves 4.4% improvement on code benchmarks while reducing autoregressive bias.
    0
    installs
80. ▌
    Distribution Matching Vae · adu2021

    Align latent distributions with arbitrary reference distributions via explicit matching constraints rather than fixed priors. DMVAE achieves gFID 3.2 on ImageNet with 64 epochs—when you need flexibility in latent representation design for image generation.
    0
    installs
81. ▌
    Emma Efficient Multimodal · adu2021

    Build a single model handling multimodal understanding, generation, and editing tasks efficiently through token compression and intelligent component sharing. EMMA-4B surpasses larger models while reducing computational burden—ideal when you need unified performance across vision and language tasks.
    0
    installs
82. ▌
    Few Step Distillation T2i · adu2021

    Systematically adapt state-of-the-art distillation methods for T2I generation. Compare sCM (stabilized Consistency Models), MeanFlow, and IMM within unified framework. sCM excels at extreme few-step regimes (52.81% GenEval at 2 steps), MeanFlow achieves superior fidelity at 4 NFEs.
    0
    installs
83. ▌
    Fluidworld Reaction Diffusion Models · adu2021

    Replace self-attention world models with reaction-diffusion PDEs as the predictive substrate. Demonstrate that aperiodic PDE dynamics achieve superior multi-step rollout stability and inherent error correction through Laplacian diffusion smoothing, while maintaining O(N) complexity and enabling autonomous corruption recovery.
    0
    installs
84. ▌
    Insight V Plus Plus Visual Reasoning · adu2021

    Extend visual reasoning to spatial-temporal sequences via two-agent reasoning+summary pipeline with ST-GRPO (temporal alignment) and J-GRPO (evaluative robustness) algorithms. Achieves +8.1% on image and +6.9% on video benchmarks through autonomous data generation and self-evolving rewards; enables continuous improvement loops for visual reasoning without human annotation.
    0
    installs
85. ▌
    Regulating AI Agents Policy Analysis · adu2021

    Understand the policy challenges of governing autonomous AI agents under existing frameworks like the EU AI Act. Identifies three primary governance failures: performance failures during execution, misuse risks from malicious deployment, and economic inequality in agent access. Argues that regulations designed for static AI systems inadequately address agent autonomy. Use when understanding regulatory implications of agent deployment, advocating for policy changes, evaluating governance readiness, or designing agent systems with compliance in mind.
    0
    installs
86. ▌
    Trajloom Dense Trajectory Generation · adu2021

    Improve dense trajectory generation by replacing absolute coordinate encoding with grid-anchor offset encoding (reduces location variance 90%→10%), adding spatiotemporal consistency regularizers to VAE (30-40× improvement in 81-frame prediction), and using boundary-anchored fine-tuning for flow models. Effective for predicting extended trajectories (81 frames vs. prior 24-frame limits) in autonomous driving and video prediction where motion coherence across time steps is critical.
    0
    installs
87. ▌
    Unigrpo Unified Visual Generation Rl · adu2021

    Replace classifier-free guidance and KL-based regularization in flow matching with velocity-based MSE regularization and GRPO for joint text-image optimization. Achieves 0.8381 TextAlign and 0.90 GenEval without CFG overhead. Works best for multimodal generation where text and image must stay coherent. Trigger: When optimizing vision-language models with flow matching and need better joint text-image policy.
    0
    installs
88. ▌
    Acon Context Compression Long Horizon · adu2021

    Compress agent interaction histories and environment observations through natural language guideline optimization, reducing token usage by 26-54% while preserving 95%+ accuracy. Use for cost/latency reduction in multi-step agent tasks.
    0
    installs
89. ▌
    Agent As A Judge Evaluation Framework · adu2021

    Transition from simple LLM-based evaluation to agentic judges that employ planning, tool-augmented verification, multi-agent collaboration, and persistent memory. Survey of sophisticated evaluation paradigms for complex, specialized, and multi-step assessment tasks across diverse domains.
    0
    installs
90. ▌
    Agentic Context Engineering Evolution · adu2021

    Enable language models to improve via context adaptation rather than weight updates. Use ACE (Agentic Context Engineering) to treat contexts as evolving playbooks that accumulate, refine, and organize strategies through modular generation, reflection, and curation processes. Achieve +10.6% agent benchmark gains and +8.6% on finance tasks using small open-source models matching production-level performance.
    0
    installs
91. ▌
    Arise Skill Evolution Hierarchical Rl · adu2021

    Build reusable skill libraries for mathematical reasoning through hierarchical RL. Maintain a high-level skills manager that summarizes successful solution traces and selects relevant strategies to condition future rollouts.
    0
    installs
92. ▌
    Attention Residuals For Depth Scaling · adu2021

    Replace uniform residual accumulation with depth-wise attention that selectively aggregates earlier layer representations. Improve gradient flow and model performance in deep architectures by learning content-dependent depth-wise selection.
    0
    installs
93. ▌
    Audio Flamingo 3 Multimodal Reasoning · adu2021

    Build fully open audio-language models supporting reasoning over speech, sound, and music with 10-minute long-form comprehension and multi-turn conversation capabilities. Use when you need to process audio modalities alongside text for complex reasoning tasks across speech recognition, sound classification, and music analysis.
    0
    installs
94. ▌
    Calligrapher Text Style Customization · adu2021

    Automates artistic typography customization through self-distilled learning and localized style injection. Generates stylized text images by encoding reference style and injecting it into diffusion denoising. Use for digital design workflows, text-based visual content creation, or applications needing artistic typography control without manual annotation.
    0
    installs
95. ▌
    Clipo Contrastive Policy Optimization · adu2021

    Augment verifiable reward RL (RLVR) with contrastive learning to generate dense auxiliary rewards. Enforce proximity among correct reasoning trajectories in embedding space while suppressing errors, amplifying invariant reasoning patterns.
    0
    installs
96. ▌
    Data Efficacy Language Model Training · adu2021

    Boost language model performance by strategically ordering training data without changing content or model size. Uses learnability-quality scoring and folding schedules to improve convergence and knowledge retention, achieving consistent gains across all model scales.
    0
    installs
97. ▌
    Embedding Space Multitoken Prediction · adu2021

    Accelerate LLM decoding by predicting multiple future tokens simultaneously using mask-token probing in embedding space, without retraining or auxiliary models.
    0
    installs
98. ▌
    Fasa Frequency Aware Sparse Attention · adu2021

    Identify dominant frequency components in RoPE embeddings to determine which attention dimensions are essential, reducing KV cache by 81% while maintaining performance. Training-free approach integrates with existing compression methods for 2.56x speedup on long-context reasoning.
    0
    installs
99. ▌
    Frugal Reasoning Short Math Solutions · adu2021

    Achieve emergent brevity in reasoning by retaining and up-weighting easy problems during RL training, implicitly regularizing solution length without explicit penalties while maintaining accuracy on hard problems.
    0
    installs
100. ▌
     Hardware Agnostic Reranker Evaluation · adu2021

     Evaluate LLM-based document rerankers using hardware-agnostic FLOPs metrics instead of latency, enabling fair comparison of ranking quality per unit of computation across different models and deployment scenarios.
     0
     installs