Measure Once, Model Everywhere: Model-Based Per-Request Resource Consumption for HTTP

2026-07-03T07:23:51Z234ef0f07ed3b615bc6973260dc8859c1038773453ca5d1187530e91acc10432
AI-alignmentAI-governanceCO2eHTTPLLM-gradingacademic-integritybenchmarkingcollective-empiricismdermatology-AIdisagreement-resolutioneducation-technologyenergy-disclosureevaluation-scarcitygenerative-AImodel-registrynginxpartial-credit-promptingper-request-metricspractice-auditingpseudo-rational-cognitionradiology-futuresscalable-oversightsociotechnical-alignmentsustainabilityuniversity-policy

What happened

This RSS bundle contains nine new arXiv papers (07/03/2026) spanning practical sustainability telemetry for web servers, empirical evaluations of LLMs in education, conceptual frameworks for AI governance and auditing, new oversight protocols for model disagreement, sectoral studies of AI adoption in clinical practice, workforce impacts in radiology, and meta-issues in AI evaluation and university policy. Key contributions: an nginx extension and offline/online model pipeline to emit per-request energy and CO2e estimates; a multi-model study showing liberal partial-credit prompting improves LL

Why it matters

A reviewed impact interpretation has not been published for this record.

Evidence and limitations

Source ID
arxiv_cs_cy
Record identifier
234ef0f07ed3b615bc6973260dc8859c1038773453ca5d1187530e91acc10432
Enrichment time
2026-07-03T07:23:51Z
AI-assisted enrichment
Yes

This record may overlap with other records. Its enrichment can be incomplete or wrong, and machine assistance was used. Validate consequential decisions against the linked source and your own environment.

Record · Measure Once, Model Everywhere: Model-Based Per-Request Resource Consumption for HTTP · Baitaphish