INDEX

Explanations

pseudocode and algorithm

np_max-act · gemini-2.0-flash

The neuron strongly activates on pseudocode control‐flow keywords (e.g. “for,” “foreach,” “if,” “repeat,” etc.), marking places where the algorithm branches or loops.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

.KeyPress

-0.07

 Astr

-0.07

しない

-0.07

 состав

-0.06

.cv

-0.06

 пласти

-0.06

.removeAttribute

-0.06

ICollectionView

-0.06

ерату

-0.06

 Thread

-0.06

POSITIVE LOGITS

 Constit

0.06

 револю

0.06

osti

0.06

 defining

0.06

forma

0.06

 Grip

0.06

 ellipse

0.06

【

0.05

 features

0.05

Activations Density 0.044%

pseudocode and algorithm

The neuron strongly activates on pseudocode control‐flow keywords (e.g. “for,” “foreach,” “if,” “repeat,” etc.), marking places where the algorithm branches or loops.

No Comments

No Known Activations