SAE-IDF matches symbols across systems using sparse autoencoder codes
On 12 tasks across six symbolic systems, SAE-IDF beats the same LLM's dense baseline by up to 45 points.
46
AI model
Gemma-2-9B
- Mentions · 7d0
- Mentions · 30d3
- All mentions3
- First seenSep 27, 2026
- Last mentioned12d ago
Timeline
3 articles · newest firstTask Context Shapes Component Roles in Language Models
Task context changes which facts shared language-model heads use and how they influence later computation.
42
When Does Correction Become Repair? Mechanistic Auditing of Internal Interventions in Tool-Using LLMs
SAKIKO finds activation steering can raise tool-use scores while corrupting many previously correct LLM decisions.
48
Appears with
- Qwen3-14Bai model · 1
- Qwen3-4Bai model · 1
- Llama-3.1-8Bai model · 1
Entities are extracted by the model from each article. First seen , last mentioned . Watching keeps it in this browser only (no account); the watchlist page and dashboard alerts use it.