Probes test spatial state in ten visual foundation models
Strong visible-scene readout does not ensure object persistence, future-feature retrieval or reliable action utility.
Stories per day · 30 days
1 RSS
Related3d-reconstruction1computer-vision1egocentric-vision1spatial-representation1uncertainty1vggt1visual-foundation-models1world-models1
10 stories, newest first · one card per story
Stories tagged #probing
Linear probe predicts VGGT pose failures the model does not report
On HardMatch, where VGGT fails on 35% of 1,000 pairs, a frozen-feature probe reaches 93.5% ROC-AUC.
Depth as Time in One-Step Generative Models
Depth-as-time analysis shows one-step diffusion models reorganize temporal denoising across network depth, enabling 16.6x compression of MeanFlow SiT-L/2.
User Model Extraction via Belief Self-Distillation
Belief Self-Distillation reads and rewrites an LLM's user beliefs, changing refusals with the request fixed.
CVE-2026-87902: Attackers Started Probing WordPress Sites Hours After the Patch
Attackers probed unauthenticated WordPress LFI CVE-2026-87902 within hours of the 7.1.2 patch.
5
A Lie Detector Test for Language Models: Reading Knowledge a Model Won't Reveal
PIR reads internal model states to detect concealed knowledge, distinguishing sandbagging from genuine ignorance across Gemma, Qwen, Llama, Mistral and Phi models.
Large Language Models Develop Belief State Geometry In-Context
Probing six open-source LLMs on HMM-generated data shows belief states are linearly decodable (R² 0.83–0.99), suggesting in-context learning approximates Bayesian prediction.
The Router Within: Eliciting Native Skill Routing from a Frozen LLM
Gavel reads native skill-routing signals from a frozen LLM's forward passes with two linear maps, beating retrieve-and-rerank pipelines by up to 21.9 points on Qwen3-32B.
Canonical Color as a Lens into Concept Decodability in Vision Encoders and VLMs
Probing study shows vision encoders make canonical color linearly decodable from grayscale images and tie it to object identity.
Encoded Early, Used Late: Where Transformers Begin to Act on an Inferred Partner's Expertise
Probing finds transformers represent an inferred dialogue partner's expertise in early layers long before it causally influences output.