INDEX

Explanations

code

np_max-act · gemini-2.0-flash

The neuron detects formatting or structural instructions (e.g. requests to “write,” “in math,” “as a paragraph,” “point form,” “process flow,” etc.).

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

 mái

-0.07

 flavour

-0.06

Ease

-0.06

'",

-0.06

 Robbins

-0.06

 Decorating

-0.06

('#

-0.06

 เค

-0.06

.Animation

-0.06

 salle

-0.06

POSITIVE LOGITS

 armed

0.07

ією

0.07

Cu

0.06

 نزد

0.06

 airports

0.06

 того

0.06

(ec

0.06

pictureBox

0.06

 나는

0.06

 aplikace

0.06

Activations Density 0.082%

code

The neuron detects formatting or structural instructions (e.g. requests to “write,” “in math,” “as a paragraph,” “point form,” “process flow,” etc.).

No Comments

No Known Activations

code

The neuron detects formatting or structural instructions (e.g. requests to “write,” “in math,” “as a paragraph,” “point form,” “process flow,” etc.).

No Comments

No Known Activations