HARP lifts Hit@1 in LLM-based recommendation by 6.9%-24.0%
Training-free dual-profile agentic communication framework improves recommendation accuracy over strong baselines with p < 1e-9 on four datasets.
Stories per day · 30 days
3 RSS
Relatedlanguage-models2agentic-communication1avatars1communication1diffusion-models1dual-profiles1fine-tuning1gaussian-splatting1gpu1gui-grounding1hidden-states1korteweg-de-vries1length-extrapolation1model-compression1model-transfer1moe-expert-pruning1mteb1multi-agent1
18 stories, newest first · one card per story
Stories tagged #training-free
RangeSWA provides training-free length extrapolation for diffusion language models
RangeSWA reduces perplexity and repetition in long-text generation using sliding-window attention and positional remapping.
Replacement operators affect training-free MoE expert pruning
Survivors gate scale choice dominates effects in pruning OLMoE, DeepSeek-V2-Lite and Moonlight models; tilted novice operator adds 1.2-1.8% parameters and improves accuracy.
AdaGround adds adaptive zoom and a verifier for GUI grounding
On Qwen3-VL, the training-free method raises accuracy 7.4 to 20.6 points on OSWorld-G and ScreenSpot-Pro.
Ariadne transfers fine-tuning effects across differently shaped Transformers
The training-free method synthesizes a native target update from activation changes instead of moving parameters.
HumanSkill-4D queries animatable Gaussian avatars without training
The training-free method beat baselines on part segmentation and 3D and 4D referring segmentation using three new human-annotated benchmarks.
Paired-residual removal improves multilingual LLM text embeddings
The inference-time method was tested on 18 languages and English bitext mining against more than 200 languages.
The Physics Is the Network: Fast, Training-Free Neural Emulation
Training-free PolyNets emulate known polynomial physics with solver-level accuracy and strong GPU ensemble scaling.
Share the Symbol, Not the Vector: Training-Free Fact Communication Across Language Models
SYNC lets heterogeneous language models share facts via word indices, reaching up to 98.3% without extra training.
Foresight: planning future perception in streaming VLMs without retraining
FORESIGHT lets frozen Qwen3-VL-8B plan future perception in streaming video without retraining.
Devils in Question Relay: Source-Conditioned Relay Steering to Mitigate Hallucinations in Audio-visual Large Language Models
SECRET, a training-free steering method, cuts source-confused hallucinations in audio-visual language models.
GeoPair: Geometry-Preserving Cross-Layer Factorization for Training-Free Transformer Compression
GeoPair compresses transformers training-free by pairing compatible layers and sharing factorized dictionaries.
Neural Spectral Capacity: Measuring and Designing Architectures from Network Specification Alone
Neural Spectral Capacity scores architectures from their specs alone and prunes LLaMA-7B to 5.7B without calibration data.
Refinement Is Inherently Editable: Training-Free Prompt-to-Prompt Image Editing with Generative Refinement Network
RefineEdit is a training-free image editing framework achieving best background preservation on PIE-Bench through generative refinement.
Reflect, Revise, Reuse: Training-Free Skill Evolution for GUI Agents
EvoSkill-GUI introduces training-free skill revision for GUI agents, boosting MobileWorld, AndroidWorld, and OSWorld success by up to 16.2%.
HarnessVLN: Unifying Training-Free Embodied Navigation through an Agent Harness
Researchers introduce HarnessVLN, a zero-shot training-free agent harness that sets new training-free SOTA on vision-language navigation benchmarks including R2R and HM3D.
Reflection-aware Generative Novel View Synthesis
Ref-GeNVS is a training-free method for reflection-consistent generative novel view synthesis that treats mirror images as two complementary views.
One Editor, Many Edits: A Unified Training-Free Framework for Diverse Video Editing
Researchers introduce EditVid, a training-free video editing framework scoring 78.16 FiVE-Acc versus 58.95 for the strongest comparable baseline.