ZPPO improves distillation to small vision-language models by using binary and negative candidate prompts plus a replay buffer for hard questions, outperforming standard distillation and GRPO on a 31-benchmark suite with largest gains at the 0.8B scale.
Visualpuz- zles: Decoupling multimodal reasoning evaluation from domain knowledge
10 Pith papers cite this work. Polarity classification is still indexing.
10
Pith papers citing it
citation-role summary
background 2
dataset 1
citation-polarity summary
years
2026 10representative citing papers
StemBind benchmark diagnoses MLLM failures in abstract visual reasoning by separating perception, rule induction, and answer selection on shared stems, finding a persistent rule-to-instance binding gap even when perception and rule are correct.
A dual local/global selection of ~100 non-agentic instances predicts four agentic benchmarks with LOOCV MAE under 4%, Spearman above 0.80, and ~85% pairwise ranking accuracy at under 1% of full agent cost.
SCOLAR fixes information gain collapse in latent visual reasoning by generating independent auxiliary visual tokens via a detransformer, extending acceptable CoT length over 30x and delivering +14.12% gains on reasoning benchmarks.
ROMA improves MLLM robustness to seen and unseen visual corruptions by +2.3-2.4% over GRPO on seven reasoning benchmarks while matching clean accuracy.
VLMs fail at exact visual path following primarily at self-intersections, where performance drops sharply after the first crossing on a controlled polyline benchmark.
MapTab introduces a 328-map, 196,800-query benchmark showing that current multimodal LLMs fall far short on multi-criteria route planning from maps-plus-tables.
Multimodal CoT improves math and science reasoning but degrades visual perception performance and exhibits a pattern of diminishing visual reflection while verbal reflection varies.
SLVR is a two-stage method that enriches region-centric latent representations with fine-grained attribute semantics and aligns them via M-GRPO across multiple queries on the same region, supported by new SLV-Set dataset and SV-QA benchmark.
citing papers explorer
-
Zone of Proximal Policy Optimization: Teacher in Prompts, Not Gradients
ZPPO improves distillation to small vision-language models by using binary and negative candidate prompts plus a replay buffer for hard questions, outperforming standard distillation and GRPO on a 31-benchmark suite with largest gains at the 0.8B scale.
-
StemBind: When MLLMs Get Lost Between Rules and Instances in Abstract Visual Reasoning
StemBind benchmark diagnoses MLLM failures in abstract visual reasoning by separating perception, rule induction, and answer selection on shared stems, finding a persistent rule-to-instance binding gap even when perception and rule are correct.
-
PACE: A Proxy for Agentic Capability Evaluation
A dual local/global selection of ~100 non-agentic instances predicts four agentic benchmarks with LOOCV MAE under 4%, Spearman above 0.80, and ~85% pairwise ranking accuracy at under 1% of full agent cost.
-
Self-Consistent Latent Reasoning: Long Latent Sequence Reasoning for Vision-Language Model
SCOLAR fixes information gain collapse in latent visual reasoning by generating independent auxiliary visual tokens via a detransformer, extending acceptable CoT length over 30x and delivering +14.12% gains on reasoning benchmarks.
-
Reinforcing Multimodal Reasoning Against Visual Degradation
ROMA improves MLLM robustness to seen and unseen visual corruptions by +2.3-2.4% over GRPO on seven reasoning benchmarks while matching clean accuracy.
-
TraversalBench: Challenging Paths to Follow for Vision Language Models
VLMs fail at exact visual path following primarily at self-intersections, where performance drops sharply after the first crossing on a controlled polyline benchmark.
-
MapTab: A Diagnostic Benchmark for Long-Horizon Multi-Criteria Multimodal Reasoning on Heterogeneous Topological Graphs
MapTab introduces a 328-map, 196,800-query benchmark showing that current multimodal LLMs fall far short on multi-criteria route planning from maps-plus-tables.
-
Look Light, Think Heavy: What Multimodal Chain-of-Thought Reasoning Can and Cannot Do
Multimodal CoT improves math and science reasoning but degrades visual perception performance and exhibits a pattern of diminishing visual reflection while verbal reflection varies.
-
Semantic-Enriched Latent Visual Reasoning
SLVR is a two-stage method that enriches region-centric latent representations with fine-grained attribute semantics and aligns them via M-GRPO across multiple queries on the same region, supported by new SLV-Set dataset and SV-QA benchmark.
- SALLIE: Generation-Free Hidden-State Detection of Jailbreaks and Prompt Injections Across Text and Vision