Storia in 2 fonti

[AINews] How to steal a Reasoning Trace

Speculative Decoding by any other name would distil as sweet

Raccontata da

latent.space

dev.to

Confronto fonti

2 prospettive sulla stessa storia

AI · summaries

latent.spaceStai leggendo8 g fa

[AINews] How to steal a Reasoning Trace

Speculative Decoding by any other name would distil as sweet

originale

dev.to6 g fa

Stealing Reasoning Traces from LLM APIs: How It Works and What to Audit

Researchers extracted encrypted reasoning blocks from Anthropic, OpenAI, Google via replay to weaker models, recovering hidden thinking in two API calls. Public logs expose 315k reconstructed reasoning blocks with 704 private artifacts including API keys, passwords—64 visible only inside blocks, missed by redaction.

Leggi questa versione → originale

Timeline cronologica

mercoledì 12 agosto 2026·latent.space
[AINews] How to steal a Reasoning Trace
Speculative Decoding by any other name would distil as sweet
venerdì 14 agosto 2026·dev.to
Stealing Reasoning Traces from LLM APIs: How It Works and What to Audit
A paper from researchers at ELLIS Institute Tübingen, the Max Planck Institute for Intelligent...