All Dev Resources
3123 resources total
Arcaide
★ 80A code exploration tool that uses multi-level call graphs to visualize and navigate code.
FedKT-CSD
★ 80Federated Knowledge Transfer via Collaborative Synthetic Data for private knowledge transfer in FL.
HIVE
★ 80Hallucination Inference and Verification Engine for studying post-hallucination reasoning in VLMs.
FlightPlanningLLMs
★ 80Code implementation for End-to-End LLM Flight Planning with RAG and multi-modal agents.
KD-4-TSC
★ 80Implementation of knowledge distillation for time series classification across FCN, Inception, and ConvTran architectures.
tsbootstrap
★ 80Distribution-free uncertainty quantification and conformal prediction for time series.
NEST
★ 80A regime-oriented Mixture-of-Experts framework for multivariate time-series forecasting.
A curated list of resources and papers regarding replication in visual diffusion models.
Implementation and experimental code for improving LLM-generated process models via RL.
PluraMath
★ 80An open-source dataset and framework for evaluating mathematical reasoning in underrepresented languages.
SHIELD
★ 80A multi-agent system designed to prevent Knowledge Debt by surfacing learning moments in AI-assisted coding.
Flint
★ 80A visualization language for AI agents.
ResonatorLM
★ 80A physics-derived alternative to attention for efficient long-context language modeling.
Psychometric study separating logical verdict from lexical and order biases in LLMs.
KAT-Coder
★ 80An agentic coding model capable of autonomous operation within executable repositories.
CanvasAgent
★ 80A multimodal agent that learns to orchestrate heterogeneous visual tools for image creation.
KVpop
★ 80Key-Value Cache Compression with Predictive Online Pruning for autoregressive decoding.
IGRPO
★ 80An adaptive tree-structured rollout approach for multi-turn LLM agents based on information gain.
AgoraSim
★ 80Hybrid agent-based modeling framework for scenario-oriented social reaction analysis.
SPLIT
★ 80A training-free detector for AI-generated and partially edited video detection.