LVLMs
PulseAugur coverage of LVLMs — every cluster mentioning LVLMs across labs, papers, and developer communities, ranked by signal.
- instance of Large Vision Language Models 90%
- instance of ScienceCast 90%
- instance of CatalyzeX 90%
- instance of alphaXiv 90%
- instance of Gotit.pub 90%
- authored by alphaXiv 70%
- used by alphaXiv 70%
- used by Gotit.pub 70%
- used by ScienceCast 70%
- used by CatalyzeX 70%
- used by DagsHub 60%
- used by Large Vision Language Models 60%
8 day(s) with sentiment data
-
New probe reveals how vision-language models ground images internally
Researchers have developed a new method to understand how large vision-language models (LVLMs) identify relevant images within interleaved text and image inputs. They introduced ICIR-MCQ, a controlled probe that analyze…
-
New methods prune visual tokens in LVLMs for efficiency
Researchers have developed new methods to prune visual tokens in large vision-language models (LVLMs), significantly reducing computational costs without substantial performance loss. The first approach, BtsPruner, iden…
-
New TARGRS framework detoxifies LLMs and LVLMs with minimal impact
Researchers have developed a new test-time framework called TARGRS to reduce toxic content generation in Large Language Models (LLMs). This method guides representation sampling by learning from transitions between toxi…
-
New benchmarks reveal limitations in MLLM visual reasoning capabilities
Researchers are developing new benchmarks and methods to evaluate the visual reasoning capabilities of multimodal large language models (MLLMs). One new benchmark, "The Prestige," uses magic tricks to challenge models' …
-
New method optimizes LVLMs for 3D-VQA with token pruning and coding
Researchers have developed TP-JSCC, a novel method for optimizing large vision-language models (LVLMs) in cloud-edge collaboration for 3D visual question answering (3D-VQA). The approach includes a Local and Global Sema…
-
Sketches reveal structural understanding in LVLMs, pinpointing grounding layers
Researchers have developed a novel method to analyze Large Vision-Language Models (LVLMs) by using sketches, which strip away photometric details to reveal structural understanding. This technique assesses how well LVLM…
-
New MAGIC method enhances vision-language model robustness against bias
Researchers have developed a new method called MAGIC (Margin-gated Contrastive Debiasing with Evidence Restoration) to improve the robustness of large vision-language models (LVLMs) against language bias during inferenc…
-
llama.cpp updates, AI code tools, hallucination mitigation research, and LLM neutrality discussed
The llama.cpp project has released updates focusing on performance and compatibility across various operating systems and hardware. Release b11977 addresses issues with the ggml-openvino integration for SCALE operations…
-
New 'Grounding IDs' concept boosts LVLM performance
Researchers have introduced a new concept called "Grounding IDs" to improve the performance of large vision-language models (LVLMs). These latent identifiers are formed when external visual cues, such as symbols or grid…
-
New framework GeoCloak enhances image geolocation privacy
Researchers have developed GeoCloak, a novel agentic framework designed to protect geolocation privacy in images. This system addresses the risks posed by large vision language models (LVLMs) that can infer locations fr…
-
New MVGraph benchmark reveals LVLM struggles with multi-view visual graph reasoning
Researchers have introduced MVGraph, a new benchmark designed to evaluate Large Vision-Language Models (LVLMs) on their ability to compose evidence from multiple interconnected visual graph views. Existing benchmarks ty…
-
New benchmarks and frameworks advance AI spatial reasoning capabilities
Researchers are developing new frameworks and benchmarks to improve spatial reasoning in AI models, particularly Large Vision-Language Models (LVLMs). Projects like Anvil focus on unified data generation and training fo…
-
New frameworks enhance LVLM spatial reasoning with soft thinking and self-distillation · 3 sources tracked
Researchers have developed two novel approaches to enhance spatial reasoning in large vision-language models (LVLMs). One method, Soft Spatial Reasoning, introduces a "soft thinking" framework that allows models to main…
-
New TestHallVQA benchmark probes LVLMs' reasoning with redundant context
Researchers have introduced TestHallVQA, a new benchmark designed to evaluate Large Vision-Language Models (LVLMs) on their ability to perform document-level reasoning, particularly in the presence of redundant or irrel…
-
New PetQA benchmark evaluates AI veterinary knowledge
Researchers have developed PetQA, a new benchmark designed to evaluate the veterinary knowledge and clinical reasoning capabilities of large language models (LLMs) and large vision-language models (LVLMs). The benchmark…
-
New synthetic dataset enhances LVLM spatial reasoning capabilities
Researchers have introduced SpatialBlock, a synthetic dataset designed to improve the spatial intelligence of Large Vision-Language Models (LVLMs). The dataset comprises 15,000 block-stacking problems that cover 3D-to-2…
-
New Cen-Prune method enhances LVLM efficiency by optimizing visual token pruning
Researchers have developed a new method called Cen-Prune to improve the efficiency of Large Vision-Language Models (LVLMs) by optimizing how visual tokens are pruned. Standard diversity-based pruning relies on cosine si…
-
New dataset Synth-JDoc boosts LVLM Japanese OCR performance
Researchers have developed Synth-JDoc, a novel dataset designed to improve the optical character recognition (OCR) capabilities of Large Vision Language Models (LVLMs), particularly for Japanese text. This dataset synth…
-
New benchmark reveals LVLMs struggle with interactive visual grounding
A new research paper introduces a framework for evaluating interactive visual grounding in large vision-language models (LVLMs). The study highlights that current LVLMs struggle with tasks requiring dialogue to refine o…
-
New benchmark and training method boost multilingual LVLM performance
Researchers have introduced PM4Bench, a new benchmark designed to evaluate the multilingual capabilities of Large Vision-Language Models (LVLMs). This benchmark utilizes a parallel corpus across 10 languages and incorpo…