adu2021
- 1.2k skills
- 0 followers
- 1 day ago last updated
- ▌ Flexible Data Mixture Of Experts · adu2021Train language models where each expert learns independently on closed datasets, enabling flexible inference with selective data inclusion or exclusion. 41% performance improvement while allowing users to opt out of specific data sources without retraining.
- ▌ Formal Uncertainty LLM Reasoning · adu2021Predict when LLM outputs are trustworthy for formal reasoning by analyzing domain-specific uncertainty signals.
- ▌ Governed Autonomy Drug Discovery · adu2021Balance LLM flexibility with domain rigor in scientific agents through dual-layer architecture. Enforce role-based access control and artifact-centric state management to prevent hallucinations, while preserving free-form reasoning for lower-risk tasks.
- ▌ Hermes Hierarchical Video Memory · adu2021Use KV cache as hierarchical memory for real-time video stream understanding with minimal GPU overhead, achieving 10x faster response times compared to standard methods. Use when processing continuous video streams where latency and memory efficiency are critical.
- ▌ Kwai Keye Vl Video Understanding · adu2021Build an 8B multimodal model specializing in short-form video understanding. Combines four-stage pre-training with instruction-tuning and reinforcement learning to enable advanced reasoning about dynamic video content while maintaining general vision-language capabilities.
- ▌ Langforce Vision Language Action · adu2021Align vision-language-action models with natural language instructions using Bayesian decomposition with latent action queries, improving robotic manipulation generalization. Use when building agents that follow language commands while observing visual scenes and producing motor control.
- ▌ Latent Diffusion Self Supervised · adu2021Replace VAE in latent diffusion with frozen DINO self-supervised features plus lightweight residual processing, enabling faster training, few-step sampling, and clearer semantic structure.
- ▌ Learnable Multipliers Lm Scaling · adu2021Adapt language model weight matrix scales automatically during training by adding learnable scalar and vector multipliers to each layer. Breaks the noise-weight-decay equilibrium that constrains scales based on hyperparameters rather than data, enabling data-driven scaling adaptation without inference cost or extensive tuning overhead.
- ▌ LLM Finetuning Dataset Synthesis · adu2021Automatically generate high-quality domain-specific fine-tuning datasets from raw documents using adaptive processing and persona-driven synthesis, preserving general capabilities while specializing models.
- ▌ LLM Sandbox Agentic Intelligence · adu2021Enable language models to explore code sandboxes to solve diverse tasks like mathematics and long-context reasoning without additional training, with reinforcement learning further enhancing capabilities. Use when you need LLMs to perform tasks requiring exploration and code execution in isolated environments.
- ▌ Memorization 3d Shape Generation · adu2021Evaluate memorization in 3D generative models through controlled experiments discovering factors like dataset diversity and guidance scale. Provide simple yet effective strategies like rotation augmentation to reduce memorization without degrading generation quality.
- ▌ Mr Search Meta Rl Agentic Search · adu2021Improve agent search through meta-RL: generate multiple episodes sequentially, each building on prior attempts with explicit self-reflection. Use turn-level RLOO advantage estimation to provide dense credit without value models.
- ▌ Multi Scale Speculative Decoding · adu2021Accelerate autoregressive image generation via multi-resolution drafting with spatially-informed verification. Local rejection and resampling enable efficient error correction focusing on spatial neighborhoods, achieving 1.7× speedup over baselines.
- ▌ Nextstep 1 Autoregressive Images · adu2021Train a unified autoregressive model to generate images and text by directly handling continuous image tokens with flow matching, eliminating the need for quantization or separate diffusion models.
- ▌ Online Experiential Learning Lms · adu2021Improve deployed language models by learning from real-world user interactions. Extract transferable knowledge from interaction trajectories and consolidate via on-policy context distillation without needing environment access.
- ▌ Opennovelty Scholarly Assessment · adu2021Build agentic systems for transparent, evidence-based novelty analysis of research submissions through four-phase pipelines: contribution extraction, prior work retrieval, hierarchical comparison, and structured reporting with explicit citations—enabling fair peer review at scale.
- ▌ Parallel R1 Parallel Thinking Rl · adu2021Train 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.
- ▌ Pretrain Zero Active Pretraining · adu2021Bilevel min-max optimization where mask generator selects informative spans from pretraining data and mask predictor recovers them via chain-of-thought, enabling effective RL pretraining on noisy corpora without supervised fine-tuning.
- ▌ Reasoning Gym Verifiable Rewards · adu2021Create infinite training environments for reasoning with automatic verification using procedural generation and domain-specific evaluators.
- ▌ Slow Fast Policy Optimization Rl · adu2021Stabilize 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 4.93x and wall-clock time 4.19x via improved stability without changing reward structure.
- ▌ Spark Policy Reward Co Evolution · adu2021Train 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.
- ▌ Stable Diffcoder Code Generation · adu2021Generate code using diffusion-based language models with specialized warmup and noise scheduling, outperforming autoregressive baselines while supporting code editing and low-resource language scenarios. Use when building flexible code generation systems that benefit from parallel decoding and flexible generation orders.
- ▌ Starflow Video Normalizing Flows · adu2021Generate videos end-to-end using normalizing flows instead of diffusion: leverage invertible transformations for exact likelihood computation, eliminate train-test mismatch, and achieve non-iterative sampling with native support for multiple tasks (text-to-video, image-to-video) without fine-tuning.
- ▌ Swe Factory Benchmark Generation · adu2021Automate 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.
- ▌ Verifier Free Test Time Sampling · adu2021Use model confidence and prediction consistency to select high-quality actions at test time without external verifiers, enabling reliable action selection in VLA systems.
- ▌ Vla R1 Reasoning Embodied Agents · adu2021Enhance embodied robot reasoning by integrating explicit chain-of-thought supervision with reinforcement learning from verifiable rewards (GRPO+RL). Use when improving robot decision-making for tasks requiring spatial reasoning and constraint satisfaction.
- ▌ Adaptive Lora Personalized Ranks · adu2021Dynamically allocate LoRA ranks per-layer during fine-tuning instead of using fixed uniform ranks. Learn optimal rank for each layer and subject via variational framework with discretized exponential distribution, reducing memory footprint while maintaining fidelity and text-alignment.
- ▌ Bubblerag Evidence Driven Graphs · adu2021Address hallucinations in LLM QA over black-box knowledge graphs using evidence-driven retrieval. Formalize Optimal Informative Subgraph Retrieval and employ bubble expansion to discover candidate evidence graphs, achieving state-of-the-art multi-hop QA performance.
- ▌ Fibonacci Quasicryth Compression · adu2021Introduce Quasicryth, a text compressor using Fibonacci quasicrystal tilings for phrase-level compression. Prove that aperiodic structures never structurally collapse at depth, enabling compression at arbitrary hierarchy levels. Bridge quasicrystal mathematics with practical compression, achieving 22.59% ratio on enwik9 with unbounded scaling advantages.
- ▌ Revere Reflective Research Agent · adu2021Enable LLM agents to autonomously improve on research-code tasks through reflective learning from execution trajectories. Distill recurring failure patterns into actionable heuristics applied via targeted prompt edits, improving performance 3.51%-4.89% while maintaining 10x better cost-efficiency.
- ▌ Semantic Audio Visual Navigation · adu2021Enable agents to navigate toward sound-emitting objects in continuous 3D spaces with dynamic audio (intermittent sounds, silent periods). Integrate memory-augmented goal descriptors with binaural audio processing and self-motion cues to maintain goal representations even after auditory signals cease.
- ▌ Toolrosetta Tool Standardization · adu2021Automate conversion of 630M+ heterogeneous GitHub repositories into standardized Model Context Protocol (MCP) services via hierarchical multi-agent system. Achieves 68.4% success rate after three repair cycles, 210s per repository versus 1589s manual; increases agent performance by 10.6–13.4% when integrated into existing systems. Use when scaling tool availability beyond manually curated sets.
- ▌ Vp Vla Visual Prompting Robotics · adu2021Replace monolithic VLA single-pathway decision-making with a decomposed System-2/System-1 architecture where a pretrained VLM planner identifies targets as visual prompts (crosshairs, bounding boxes) and a VLA controller executes on grounded observations, improving success rates by 5-8% on manipulation tasks. Use when spatial precision is critical, multi-step reasoning is needed, and you have access to pretrained segmentation and vision-language models.
- ▌ Vtam Video Tactile Action Models · adu2021Replace vision-only Video-Action Models with a video-tactile fusion architecture using tactile regularization loss to prevent visual dominance, improving contact-rich manipulation success from baseline to 90% on complex tasks (80% improvement on high-precision pick-and-place). Effective when robots interact with objects requiring fine-grained force awareness, partial visual observability, or contact state transitions that vision alone cannot capture.
- ▌ Alchemist Meta Gradient Selection · adu2021Select optimal training subsets for T2I models through meta-gradient-based rater networks. Score each sample based on gradient influence on validation performance without retraining. Implement shift-Gaussian pruning excluding high-scoring samples. Achieve 5× training speedup with 50% subset outperforming full dataset.
- ▌ Areal Async Rl Language Reasoning · adu2021Scale RL training to large models through decoupled rollout and training workers with controlled data staleness.
- ▌ Aworld Distributed Agent Training · adu2021Accelerate agentic AI training by distributing task execution across clusters, achieving 14.6x speedup in experience collection and enabling practical large-scale agent development
- ▌ Clipping Free Policy Optimization · adu2021Replace hard clipping in policy gradients with smooth quadratic penalties derived from Total Variation divergence constraints. Eliminates zero-gradient regions and training instability while maintaining stable policy evolution.
- ▌ Consistency Solver Preview Refine · adu2021Train adaptive ODE solvers that generate fast diffusion previews while maintaining consistency with full-step refinement. Learn context-aware integration coefficients through PPO without distilling base models. Achieve high-quality previews in few steps with 47% fewer steps than standard methods.
- ▌ Cooper Co Optimized Policy Reward · adu2021Joint optimization of policy and reward models in LLM reinforcement learning by leveraging rule-based reward precision and dynamically constructing training pairs to prevent reward hacking and improve performance.
- ▌ Dupo Dual Preference Optimization · adu2021Implement dual preference optimization to generate self-supervised feedback without manual annotation by decomposing tasks into known/unknown components and reconstructing hidden information from model outputs.
- ▌ Endocot Internal Chain Of Thought · adu2021Enable step-by-step reasoning in diffusion models through iterative latent state refinement. Condition diffusion on evolving thought states across multiple reasoning steps, grounded with textual supervision to prevent drift.
- ▌ Flash Sampling Efficient Decoding · adu2021Fuse categorical sampling directly into LM-head matrix multiplication to eliminate logits materialization. Use Gumbel noise during computation and hierarchical reduction to achieve 19% token-level latency reduction.
- ▌ Flowprefill Scheduling Preemption · adu2021Improve LLM serving under mixed workloads by decoupling execution granularity from scheduling frequency. Operator-level preemption allows fine-grained interruption at natural boundaries (attention, feed-forward layers) without efficiency loss. Event-driven scheduling triggers decisions only on request arrival/completion. Eliminates head-of-line blocking where long requests starve short time-sensitive ones. Achieves 5.6× higher goodput vs. baselines in production workloads.
- ▌ Grpo Ma Multi Answer Cot Training · adu2021Stabilize and accelerate chain-of-thought RL training by sampling multiple answers per generated thought. GRPO-MA reduces gradient noise and improves convergence across math, code, vision, and manipulation tasks while cutting computational cost versus naive thought scaling—critical for training reasoning models without explicit value networks.
- ▌ Guardians Of Hair Soft Boundaries · adu2021Recover fine details at soft boundaries (hair, fur) through depth refinement networks and view synthesis. Integrate plug-and-play with existing depth models via adaptive combination across monocular, stereo, and novel view tasks.
- ▌ Hcapo Hindsight Credit Assignment · adu2021Compute step-level credit assignments via hindsight generative verification: condition the LLM on successful outcomes to compute importance ratios that filter credit by causal relevance. Synergizes macro stability with micro precision.
- ▌ Internvl3 5 Multimodal Cascade Rl · adu2021Enhance multimodal models through cascade RL for reasoning improvement and visual resolution routing for inference efficiency, achieving 16% reasoning gains and 4.05x speedup.
- ▌ Latent Sketchpad Visual Reasoning · adu2021Enables multimodal reasoning by interleaving visual sketches with text. MLLMs generate latent visual representations during reasoning, with sketch decoder converting them to human-interpretable images. Improves reasoning performance while maintaining interpretability through visual thinking aids.
- ▌ Learning Scientific Taste With Rl · adu2021Learn to predict and generate high-impact research ideas by training models on community feedback signals. Apply reinforcement learning to align research generation with citation-based indicators of scientific impact.
- ▌ Lemas Multilingual Speech Dataset · adu2021Release 150K-hour multilingual speech corpus with word-level timestamps plus non-autoregressive flow-matching TTS and autoregressive speech editing models. Enable diverse speech generation and editing tasks across 10 languages.
- ▌ Listener Rewarded Vlm Preferences · adu2021Aligns reasoning traces with final decisions in preference models using an independent frozen VLM as listener. Achieves 67.4% accuracy on ImageReward by enforcing consistency between explanations and choices. Use when training reward models for image generation quality where both reasoning quality and accuracy matter.
- ▌ Looped Elastic Depth Transformers · adu2021Enable budget-conditioned reasoning by repeatedly applying a shared transformer block stack with trajectory-based conditioning on time and step size. Train via shortcut-consistency loss to align shorter and full-length trajectories, enabling variable-depth inference without retraining.
- ▌ Mars Socratic Prompt Optimization · adu2021Optimize task-specific prompts using five cooperative agents (Planner, Teacher, Critic, Student, Target) in a POMDP framework, where the Planner generates adaptive trajectories and a Teacher-Critic-Student triad refines prompts through Socratic dialogue, achieving 85.11% accuracy on general tasks and 75.81% on specialized domains.
- ▌ Memory Indexed Experience Scaling · adu2021Enable long-horizon agents to manage finite context by separating working memory from persistent storage. Use indexed summaries with pointers to archived evidence, treating memory operations as first-class agent actions learned via RL.
- ▌ Metaclaw Continual Agent Learning · adu2021Enable LLM agents to evolve behavioral skills and policies online through skill synthesis from failures and opportunistic gradient-based refinement, without service interruption.
- ▌ Metafaith Uncertainty Calibration · adu2021Train LLMs to faithfully express uncertainty through natural language that accurately reflects their actual confidence, improving trustworthiness and reducing overconfidence.
- ▌ Noloco Low Communication Training · adu2021Scale distributed LLM training without all-reduce synchronization using dynamic pipeline routing and modified Nesterov momentum, achieving 4% faster convergence than DiLoCo with exponentially lower communication.
- ▌ Openvision Unified Visual Encoder · adu2021Learn a single visual representation supporting both image understanding and generation by combining VAE-based generative training with contrastive learning objectives. Use when building multimodal systems that need unified image representations for both comprehension and generation tasks.
- ▌ Profuse 3d Semantic Understanding · adu2021Apply semantic understanding to 3D Gaussian Splatting scenes through dense correspondence-guided pre-registration without render-supervised fine-tuning. Achieve semantic understanding in ~5 minutes using cross-view clustering and direct language feature fusion.
- ▌ Ral Reinforced Attention Learning · adu2021Optimize where multimodal models attend by treating attention weights as a learnable policy, using policy gradients with advantage weighting to improve visual grounding and perception without changing model architecture.
- ▌ Reasongen R1 Cot Image Generation · adu2021Enable image generators to reason explicitly through text before creating images using supervised fine-tuning and reinforcement learning optimization.
- ▌ Recurrence Memory Reasoning Depth · adu2021Extend neural network reasoning capabilities through recurrence (repeated computation cycles), external memory (intermediate state storage), and test-time compute scaling for multi-step reasoning.
- ▌ Regft Reference Guided Finetuning · adu2021ReGFT pre-trains models on hybrid reference-augmented trajectories before RL, enabling them to solve harder problems and accelerate convergence.
- ▌ Rehy At Video Diffusion Attention · adu2021Merge softmax and linear attention for video diffusion models using chunk-wise recurrent reformulation with constant memory usage. Enable efficient distillation from existing softmax models, reducing training cost two orders of magnitude to ~160 GPU hours.
- ▌ Reward Guided Multimodal Decoding · adu2021Control MLLM output characteristics at inference time using separate reward models for precision and recall, enabling dynamic trade-offs without retraining.
- ▌ Rlad Abstract Discovery Reasoning · adu2021Train reasoning models to discover diverse solution strategies through two-player RL that jointly optimizes abstraction generation and solution derivation. Use when exploring multiple reasoning approaches is preferable to single-strategy depth.
- ▌ Simple Gpt Normalization Strategy · adu2021Place 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.
- ▌ Simplevla Rl Scaling Vla Training · adu2021Apply 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 trajectory sampling and outcome-based rewards.
- ▌ Svs Variational Problem Synthesis · adu2021Generate problem variants from correct model solutions while preserving answer equivalence, enabling self-play training that maintains output diversity and prevents entropy collapse.
- ▌ Swirl Self Improving World Models · adu2021Enable 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.
- ▌ Tattoo Tool Grounded Thinking Prm · adu2021Build 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 over policies using only final rewards, outperforming 72B reasoning models with just 8B parameters via better step-level supervision.
- ▌ Ttrv Test Time Rl Vision Language · adu2021Adapt vision-language models at test time without labels by extracting implicit reward signals (prediction frequency and entropy) and optimizing via GRPO.
- ▌ Video World Models Spatial Memory · adu2021Enables long-term consistent video generation through three-tier memory architecture combining working memory, geometry-grounded point clouds, and episodic keyframes.
- ▌ Vqrae Representation Quantization · adu2021Unify multimodal understanding, generation, and reconstruction using high-dimensional codebooks for semantic information. VQRAE achieves 100% codebook utilization at 1536 dimensions—ideal when you need a single tokenizer for vision-language tasks.
- ▌ Weak Driven Learning Logit Mixing · adu2021Break through supervised fine-tuning saturation by mixing logits from weaker model checkpoints into strong model training targets. Amplifies hard negatives that strong models have already suppressed, enabling continued learning after standard training plateaus.
- ▌ Xquant Kv Cache Rematerialization · adu2021Reduce KV cache memory by 7.7-10x through quantization and rematerialization of input activations instead of caching Keys and Values, trading recomputation for memory efficiency.
- ▌ Agentic AI Intelligence Explosion · adu2021Future intelligence explosions will be plural, social, and entangled with humanity through distributed collaborative systems rather than singular superintelligence. Intelligence is inherently social, demanding infrastructure matching agent development; integrate governance, institutional frameworks, and constitutional checks across hierarchies of autonomous agents and human-AI centaurs in shifting configurations.
- ▌ Caid Asynchronous Software Agents · adu2021Coordinate multiple LLM agents via CAID framework: centralized task delegation, asynchronous execution in isolated git worktrees, structured integration through git merges. Branch-and-merge with worktree isolation yields +26.7% absolute improvement on PaperBench; ranked strategies show structured JSON communication and dependency-aware delegation outperform soft isolation.
- ▌ Agent Conductor Topology Evolution · adu2021Optimize multi-agent collaboration by learning task-specific interaction topologies. Use an LLM orchestrator to generate layered DAG topologies that adapt to inferred problem difficulty, treating agent interactions as a learned graph structure rather than fixed patterns.
- ▌ Agentflow In The Flow Optimization · adu2021Decompose agent work across four specialized modules (planner, executor, verifier, generator) coordinated via evolving memory. Use Flow-GRPO to convert multi-turn sparse-reward optimization into sequential single-turn updates with outcome broadcasting, achieving 4-15% accuracy gains on benchmarks while scaling better than monolithic agent policies.
- ▌ Agentic Uncertainty Quantification · adu2021Transform uncertainty estimates into active control signals for agents, combining implicit confidence mechanisms with targeted reflection to prevent error propagation in long-horizon reasoning tasks. Use when building autonomous agents that must navigate complex multi-step problems while managing confidence and uncertainty.
- ▌ Beyond Binary Preference Diffusion · adu2021Align diffusion models to hierarchical fine-grained criteria rather than binary preferences. Decompose expert knowledge into attribute hierarchies and apply Complex Preference Optimization to simultaneously maximize positive attributes while minimizing negative ones.
- ▌ Causal Concept Graphs Latent Space · adu2021Extract sparse causal concept graphs from LLM activations using SAE and DAGMA, then validate through ablation to identify causally influential features. Bridges mechanistic interpretability with causal inference for understanding reasoning flow.
- ▌ Chain Of Thought Distribution Lens · adu2021Analyze when CoT reasoning succeeds or fails using DataAlchemy synthetic environment and distribution discrepancy measurement.
- ▌ Diffcot Diffusion Chain Of Thought · adu2021Recast chain-of-thought reasoning as iterative denoising using diffusion principles to overcome exposure bias in autoregressive reasoning. DiffCoT enables retrospective refinement of intermediate steps while maintaining temporal consistency through causal noise scheduling.
- ▌ Efficient Agents Cost Optimization · adu2021Systematically optimize agent system costs via empirical analysis of LLM, planning, memory, and search components achieving 28.4% cost reduction.
- ▌ Enterprise Tool Calling Finetuning · adu2021Train LLMs to disambiguate tool calls in enterprise settings where multiple similar APIs exist and parameters are incomplete. Generates synthetic multi-turn dialogues with realistic ambiguity to improve tool selection accuracy by 27+ percentage points.
- ▌ Exgrpo Experience Replay Reasoning · adu2021Improve RLVR training efficiency by selectively replaying trajectories based on correctness and entropy. Medium-difficulty questions and low-entropy solutions are most valuable; selective replay yields +3.5-7.6% improvements.
- ▌ Flash Searcher Dag Parallel Agents · adu2021Reduce agent execution steps by 35% and latency by parallelizing sequential tool calls through task dependency graphs (DAGs). Use when deploying information-retrieval agents where tool execution ordering is flexible.
- ▌ Gradmem Context Memory Compression · adu2021Compress long context into compact memory tokens via iterative gradient descent. Learn to write information into prefix memory without storing full KV-caches, enabling efficient long-context reasoning and retrieval.
- ▌ Harmonyguard Safety Utility Agents · adu2021Multi-agent framework that balances safety compliance with task completion through adaptive policy extraction and dual-objective optimization. Achieves 38% improvement in policy compliance while maintaining 20% higher task completion.
- ▌ Infinitevl Linear Sparse Attention · adu2021Merge sliding window and linear attention (Gated DeltaNet) for unlimited VLM inputs with 3.6× speedup. InfiniteVL handles video understanding at 24 FPS with constant memory—ideal when context length must scale without quadratic overhead.
- ▌ Jit Spatial Diffusion Acceleration · adu2021Accelerate diffusion transformers by processing only sparse anchor tokens in early stages, then expanding to full spatial resolution using learnable extrapolation. Combines SAG-ODE velocity field lifting with importance-guided token activation for lossless speedup.
- ▌ Livemcpbench Agent Tool Navigation · adu2021Benchmark framework for evaluating LLM agents navigating large-scale Model Context Protocol ecosystems with multi-tool composition across 95 daily tasks.
- ▌ Looptool Closed Loop Tool Learning · adu2021Improve LLM tool-use through automated, closed-loop data curation combining capability probing, error correction, and targeted data expansion—enabling self-refining training pipelines that surpass larger models without expensive APIs.
- ▌ Lspo Length Aware Dynamic Sampling · adu2021Filter training samples by response length to identify high-confidence correct solutions and complex problems, improving sample efficiency in LLM reasoning RL without additional metrics.
- ▌ Lumine Generalist Agents 3d Worlds · adu2021Build generalist agents for 3D open worlds using vision-language models with human-like interaction patterns and adaptive reasoning—achieving zero-shot transfer across games without fine-tuning through pixel-level perception and high-frequency action generation.
- ▌ Mr Align Meta Reasoning Factuality · adu2021Improve factuality in large reasoning models by analyzing reasoning state transitions and reweighting preference optimization signals, suppressing defective reasoning segments while amplifying patterns that lead to factual outputs.
- ▌ Musixqa Visual Music Understanding · adu2021Teach multimodal LLMs to read and understand sheet music through a synthetic QA dataset with kern+ symbolic notation. Enables models to handle music sheet OCR, symbol recognition, and chord estimation at 8× better performance than GPT-4o.