Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs

Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs | Signal Canvas | ScienceToStartup

Page Freshness

Signal Canvas proof surface

Canonical route: /signal-canvas/sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms

stale

Proof freshness: stale
Proof status: unverified
Display score: 6/10
Last proof check: 2026-05-11
Score updated: 2026-05-11
Score fresh until: 2026-06-10
References: 0
Source count: 4
Coverage: 83%

This page is showing the last landed evidence receipt and score bundle because the latest proof data is outside the freshness window.

Agent Handoff

Canonical ID sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms | Route /signal-canvas/sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms

REST example

curl https://sciencetostartup.com/api/v1/agent-handoff/signal-canvas/sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms

MCP example

{
  "tool": "search_signal_canvas",
  "arguments": {
    "mode": "paper",
    "paper_ref": "sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms",
    "query_text": "Summarize Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs"
  }
}

source_context

{
  "surface": "signal_canvas",
  "mode": "paper",
  "query": "Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs",
  "normalized_query": "2605.07447",
  "route": "/signal-canvas/sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms",
  "paper_ref": "sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms",
  "topic_slug": null,
  "benchmark_ref": null,
  "dataset_ref": null
}

Evidence Receipt

Route status: building

Claims: 0

References: Pending verification

Proof: Verification pending

Freshness state: computing

Source paper: Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs

PDF: https://arxiv.org/pdf/2605.07447v1

Repository: https://github.com/conan1024hao/SAEgis

Source count: 4

Coverage: 83%

Last proof check: 2026-05-11T20:36:04.496Z

Signal Canvas receipt window

Watch and verify: Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs

/buildability/sparse-autoencoders-as-plug-and-play-firewalls-for-adversarial-attack-detection-in-vlms

Watchwatch

Subject: Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs

Verdict

Watch

Sparse Autoencoders as Plug-and-Play Firewalls for Adversarial Attack Detection in VLMs

Use Signal Canvas as the narrative proof surface

Use this Signal Canvas via API or MCP

Signal Canvas proof surface