01Active
Capability evaluations
Task-grounded tests for autonomy, scientific reasoning, tool use, and long-horizon execution.
- Agentic behavior
- Scientific uplift
- Evaluation integrity
GPT-Rosalind CERTIFIED
Independent AI research
We study what models can do before capability outruns control.
Boundary Signal evaluates frontier AI systems across capability, alignment, biosecurity, and adversarial robustness—then turns evidence into decisions.
Observed
Emerging
Controlled
00 / Thesis
The gap we work in
Power is measurable. Safety has to be, too.
Frontier models are becoming capable in domains where failure is consequential and conventional benchmarks are thin. We build evaluations that expose the boundary: what a system can accomplish, where safeguards break, and which interventions hold.
- Evidence
- Reproducible over rhetorical
- Threat model
- Realistic over theatrical
- Disclosure
- Useful without enabling misuse
01 / Research
Four connected programs
Research at the capability–control boundary.
We pair empirical measurement with operational threat modeling so results are useful to model builders, auditors, and public-interest institutions.
02Active
Alignment science
Behavioral and mechanistic studies of control, oversight, deception, and goal-directed behavior.
- Control evaluations
- Scalable oversight
- Model organisms
03Controlled access
AI & biosecurity
Measurements of biological problem-solving uplift, safeguard reliability, and misuse pathways.
- Dual-use uplift
- Safeguard stress tests
- Risk mitigations
04Active
Adversarial testing
Structured red teaming that maps failure modes without turning findings into a misuse manual.
- Attack surfaces
- Jailbreak resilience
- Defense validation
02 / Method
From concern to evidence
Test the claim.
Map the boundary.
Change the decision.
Every project begins with a decision someone needs to make—not a benchmark looking for a purpose. Our methods are designed around consequence, reproducibility, and clear uncertainty.
-
01
Frame
Threat model
Define the actor, access, objective, constraints, and real-world consequence.
-
02
Measure
Measurement design
Build elicitation-aware tasks with controls, baselines, and independent review.
-
03
Challenge
Adversarial validation
Stress the protocol and the safeguard—not only the model under ideal conditions.
-
04
Act
Decision translation
Report findings, uncertainty, mitigations, and the evidence that would change the conclusion.
03 / Briefs
Open work
Research notes and operating standards.
Responsibility
Useful, not enabling
Some findings should travel with friction.
Dual-use research creates an obligation beyond ordinary publication. We separate public evidence from operational detail, coordinate disclosure, and use controlled access when release could materially lower the barrier to harm.
Read our standard ↗04 / About
Boundary Signal Research
Independent by design.
Collaborative by necessity.
Boundary Signal is an independent AI research company focused on the empirical gap between model capability and reliable control. We work across disciplines because consequential AI risk does not respect disciplinary borders.
We welcome research collaborations, evaluation partnerships, and careful criticism from people working on the same hard problems.
01 / Independence
02 / Scientific rigor
03 / Responsible disclosure
05 / Contact
Start with the question