CommunityJul 18, 20262026-07-184 minVerified
Why LLM Refusal Behavior Differs Across Safety Boundaries
How model-specific refusal thresholds and context handling shape overrefusal, safety boundaries, and response quality.
CommunityJul 6, 20262026-07-064 minVerified
Designing Guardrails for Agentic LLM Execution
As agentic LLMs move from answering to acting, permissions, approvals, and safety design matter more than benchmarks.
CommunityJul 6, 20262026-07-064 minVerified
What Defines Success In Home Cooking Humanoids
Home cooking humanoids should be judged by task success, time, safety, and cost, not human-like appearance.
SourceJul 4, 20262026-07-044 minVerified
Open-Weight LLM Safety Beyond Release-Time Alignment
Open-weight LLM safety should be judged not only at release, but by how easily fine-tuning can weaken safeguards later.
SourceJun 12, 20262026-06-124 minVerified
Reframing Shielded RL as Design-Time Structure Analysis
A concise look at shielded RL reinterpreted as a design-time tool for structural safety analysis, not runtime blocking.
CommunityMar 8, 20262026-03-085 minVerified
Why Custom Instructions Personas Drift Under Hierarchy
Model Spec’s chain of command can override custom instructions, causing persona and reasoning drift. Design priorities, exceptions, and fallbacks to improve reproducibility.
SourceMar 7, 20262026-03-074 minVerified
Gating Robot Autonomy Using Deep Perception Uncertainty Signals
SPIRIT uses deep perception uncertainty to gate shared autonomy, switching between semi-autonomous manipulation and haptic teleoperation.
CommunityMar 4, 20262026-03-044 minVerified
How Conversational AI Design Shapes Intimacy And Trust
Examines how warmth, memory, and consistency in conversational AI affect intimacy, trust, and safety evaluation criteria.
CommunityMar 1, 20262026-03-014 minVerified
Why Tiny Prompt Changes Can Break Robot Safety
How small prompt shifts can amplify into risky robot actions, and why alignment alone can’t guarantee physical safety.
CommunityMar 1, 20262026-03-015 minVerified
Validating Failure Modes in Vision-Agent Robotics Systems
In high-risk deployments, prioritize uncertainty, false positives/negatives, and closed-loop failure propagation over single-model scores.
CommunityFeb 25, 20262026-02-254 minVerified
Tracing Output Drift With Snapshots, Seeds, And Safety
Even with the same model alias, outputs can shift due to snapshot routing, safety behaviors, and sampling settings. Use logs and regression tests to isolate causes.
CommunityFeb 16, 20262026-02-164 minVerified
Designing Boundaries for Relationship Tests in AI Chats
How to handle relationship-test prompts in AI chats: set refusal boundaries with Safe Complete, document branching rules, and validate via evaluation.
NewsJan 27, 20262026-01-272 minVerified
California And EU Launch Investigations Into xAI Grok Safety
Explores legal investigations into xAI's Grok and the shift toward mandatory AI safety standards in California and the EU.