Observation: Astra uses filler tokens far more effectively than other models
scaling01 · x · 2026-09-26
scaling01, who has been tracking looping behavior, observes that Google's Astra appears to use filler tokens much more effectively than other models — it can leverage not only user-provided filler tokens but also filler tokens it generates itself. The observation hints at differentiated internal computation strategies and warrants further verification.
More from Models
- Micius-Flash: 27B Multimodal Model for Hardware and Semiconductor Agents — zhiman-ai · 2026-11-11
- Baidu says RSI auto-optimized ERNIE inference, 6.24x decode speedup on Kunxin P800 — xiaohu · 2026-11-11
- Baidu's ECHO: multi-token prediction reusing the last layer as a drafter — xiaohu · 2026-11-11
- ERNIE 5.5 Flash Preview: 280B-param multi-router MoE activating only 14B per token — xiaohu · 2026-11-11
- ERNIE 5.5 recreates the Pelican test: 117-second animation drawn frame-by-frame in Python — xiaohu · 2026-11-11
- 35,000+ Grok-built apps exposed patient data, plaintext passwords, and sales pipeline — The Decoder · 2026-11-11