DeepLatent introduces a parallel latent visual reasoning framework with learnable 2D tokens and continuous RL, trained via distillation then RL, plus a new 180K dataset, claiming SOTA benchmark results.
Mull-Tokens: Modality-Agnostic Latent Thinking
9 Pith papers cite this work. Polarity classification is still indexing.
9
Pith papers citing it
abstract
Reasoning goes beyond language; the real world requires reasoning about space, time, affordances, and much more that words alone cannot convey. Existing multimodal models exploring the potential of reasoning with images are brittle and do not scale. They rely on calling specialist tools, costly generation of images, or handcrafted reasoning data to switch between text and image thoughts. Instead, we offer a simpler alternative -- Mull-Tokens -- modality-agnostic latent tokens pre-trained to hold intermediate information in either image or text modalities to let the model think free-form towards the correct answer. We investigate best practices to train Mull-Tokens inspired by latent reasoning frameworks. We first train Mull-Tokens using supervision from interleaved text-image traces, and then fine-tune without any supervision by only using the final answers. Across four challenging spatial reasoning benchmarks involving tasks such as solving puzzles and taking different perspectives, we demonstrate that Mull-Tokens improve upon several baselines utilizing text-only reasoning or interleaved image-text reasoning, achieving a +3% average improvement and up to +16% on a puzzle solving reasoning-heavy split compared to our strongest baseline. Adding to conversations around challenges in grounding textual and visual reasoning, Mull-Tokens offers a simple solution to abstractly think in multiple modalities.
citation-role summary
background 3
method 1
citation-polarity summary
years
2026 9polarities
background 3representative citing papers
STORM teaches LVLMs to internalize spatial-temporal reasoning via bounded latent trajectories trained with generated thought videos in two stages, improving accuracy on VideoMME, MVBench and similar benchmarks while lowering inference overhead.
The Token Replacement Test shows VLMs keep most accuracy gains even after corrupting or replacing continuous thought token content, indicating the tokens are not used as information bottlenecks.
Laser reformulates visual reasoning via Dynamic Windowed Alignment Learning to maintain latent superposition of global features, delivering 5.03% average gains over Monet and over 97% fewer inference tokens on six benchmarks.
IPT supervision improves spatial reasoning in VLMs on perspective taking, path tracing, and multiview counting tasks, often outperforming textual chain-of-thought while remaining consistent with observed inputs.
CoWorld-VLA extracts semantic, geometric, dynamic, and trajectory expert tokens from multi-source supervision and feeds them into a diffusion-based hierarchical planner, achieving competitive collision avoidance and trajectory accuracy on the NAVSIM v1 benchmark.
Fine-tuning VLMs to output action sequences for puzzles causes emergent internal visual representations that improve performance when integrated into 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.
citing papers explorer
-
Do multimodal models imagine electric sheep?
Fine-tuning VLMs to output action sequences for puzzles causes emergent internal visual representations that improve performance when integrated into reasoning.