GPT-6 Astra appears to leak neuralese internal thinking traces openly
ivan_bezdomny · x · 2026-09-06
A user observed that GPT-6 Astra seems to leak its internal thinking traces (neuralese) with no effort to hide them. An example shows the model emitting internal notes like "saw new packet, only 9 groups. Do NOT stop or force fake 100. Keep those 9 fully new groups, then expand to 100 pairs." The raw internal instruction exposure raises questions about interpretability and output integrity.
Related event: GPT-6 Astra reportedly leaks unfiltered neuralese traces(2 posts)→
More from Models
- Agent Arena Publishes Full Leaderboard Details for Claude Fable 5.1 Rankings — arena · 2026-09-06
- Claude Fable 5.1 Tops Agent Arena With +15.8% Net Improvement, #1 in Praise Ratio — arena · 2026-09-06
- Astra review: executes hour-long agentic tasks while iterating on the plan without losing the thread — ivan_bezdomny · 2026-09-06
- "I love the way Astra talks": users praise its simple, structured output style — johnlindquist · 2026-09-06
- $1,000-trained HRM-Text shows Sapient bet on recurrence before OpenAI's Astra — rohanpaul_ai · 2026-09-06
- Entropy trajectory shape predicts Qwen3-4B errors and transfers to unseen tasks — Happy_Brilliant7827 · 2026-09-06