Likely AI-generated. Pangram flags 100% of this text as AI-written.
2
E-Audit & VLM-TRACE: Sequential Safety Auditing for Multimodal Models
1
ProposalGrant
Closes December 24th, 2026$0raised
$5,500minimum funding
$20,000funding goal
Offer to donate
40 daysleft to contribute
You're pledging to donate if the project hits its minimum goal and gets approved. If not, your funds will be returned.
$20,000 left to reach the funding goal.
Project summary
Problem: Static VLM safety benchmarks fail under small edits/shortcuts; fixed-length sweeps waste API calls.
Built: Open-source VLM-TRACE for VLM provenance and perturbation testing.
Plan (Jan–Jun 2027): Combine VLM-TRACE with my IEEE ICICAIEECE 2027 anytime-valid Innovation E-Detector (alpha <= 0.05).
Output: pip install e-audit library + UK AISI inspect_ai plug-in to stop eval runs early and catch shortcut exploitation
What are this project's goals? How will you achieve them?
Jan 2027: pip install e-audit library using IEEE ICICAIEECE 2027 Innovation E-Detector (alpha <= 0.05).
Feb–Mar 2027: Run VLM-TRACE on 4 open VLMs + Indic shifts (Malayalam, Tamil, Hindi) on RunPod GPUs.
Mar 2027: Ship inspect_ai solver and scorer plug-in.
Apr–Jun 2027: Frontier API sweeps (OpenAI, Anthropic, Google), red-team spot checks, .eval logs, arXiv preprint.
How will you know it worked?
40% fewer API samples vs. fixed-length runs at alpha <= 0.05.
MIT-licensed e-audit + VLM-TRACE repos, passing CI, inspect_evals PR.
Public .eval logs and tables on Hugging Face Hub.
Preprint with results on arXiv / NeurIPS 2027
How will this funding be used?
Researcher stipend: $10,200 ($1,700/mo x 6 mos, Trivandrum, India; includes retroactive VLM-TRACE credit).
Frontier VLM/LLM APIs & harness tooling: $3,300 ($550/mo x 6 mos, invoiced).
Contingent IIT Bombay mentorship fee: $2,000 (drawn only with official invoice; otherwise unspent or moved to APIs).
Red-team checks, hosting, workshop travel: $1,900.
RunPod GPU compute: $1,800 ($300/mo x 6 mos, invoiced).
Wire/FX fees & buffer: $800.
Who is on your team? What's your track record on similar projects?
Johan Shiju Varughese — solo PI, final-year B.Tech CS (AI), MBCET, Trivandrum, India.
Built VLM-TRACE; merged PR #3 (BF16 fix) in UBC TEA Lab's vlm-probe.
First author, IEEE ICICAIEECE-2027 (Innovation E-Detector); Zenodo null-result stress test.
Chair, IET MBCET Student Chapter; panel speaker, IET Festival of Volunteering (London, Sept 2026).
Mentorship: IIT Bombay outreach unconfirmed/contingent; applying to SPAR Spring 2027 & AI Safety India.
What are the most likely causes and outcomes if this project fails?
Heavy-tailed VLM noise drops sample savings < 20% -> publish negative results, shift Months 4–6 to shortcut benchmark.
IIT Bombay mentorship falls through -> solo execution unaffected, $2,000 stays unspent.
Low adoption -> native inspect_ai CLI plug-in.
How much money have you raised in the last 12 months, and from where?
$0 USD.