Researcher: activation steering is crude, we must speak the geometry inside models
cephaloform · x · 2026-09-24
- In a thread about steering breaking models, a researcher argues activation steering remains very crude.
- Their claim: the field should start interfacing with the geometric machines inside models "in Their language"; repeated evidence shows linear separation isn't the whole picture, yet methods haven't updated accordingly.
More from Research
- Semantic operators: LLM data processing at scale needs full-stack rethink — CShorten30 · 2026-09-24
- Jev processes 100k rows for $2.50 in under 60s, demo now public — CShorten30 · 2026-09-24
- CLM-8B hits SOTA 81.6% on DeepSWE with light finetuning, up to 9x faster inference — anshulkundaje · 2026-09-24
- SpeakerMem-R1 Tops EverMemBench at 62.33% with Dual-Track Multi-Party Dialogue Memory — zju · 2026-09-24
- CMU's WhatWorkedBench Measures How Well AI Research Agents Understand Their Experiments — CarnegieMellonU · 2026-09-24
- Tencent's RewardVerse uses rubric-guided optimization to fix video reward model drift — tencent · 2026-09-24