Visualizing circuit traces of a base model across attention and MLP layers
Sauers_ · x · 2026-10-14
A circuit-tracing visualization of a base model, showing token-level attention and MLP contributions across layers L0–L3 for the phrase "structure involved the creation of", with node and edge weights denoting bit-level ablation impact.
More from Research
- ECDSA.fail quantum challenge hits 893 qubits x 2.2M Toffolis, 25% cheaper than Google's best — jedisct1 · 2026-10-14
- NTU's COAP: code replaces LLMs in the robot control loop, hitting 75% on 42 bimanual tasks — liuziwei7 · 2026-10-14
- Diffusion LMs benchmarked: 2x faster than Haiku but 60% vs 99% pass on hard coding tasks — techNmak · 2026-10-14
- Mathematicians say rumors OpenAI's Navier-Stokes proof stole two mathematicians' work may be overblown — analisereal · 2026-10-14
- Training 54 Model Organisms Shows Interpretability Audit Scores May Be a Lottery — gsarti_ · 2026-10-14
- DoubtLess cuts Qwen3-32B chain-of-thought by 31% training-free with no accuracy loss — Minghao Luo · 2026-10-14