INDEX

Explanations

weight loss

np_max-act · gemini-2.0-flash

The neuron predominantly activates on self-referential, first-person words and constructions (e.g. “I,” “me,” “want,” “make”).

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

워

-0.07

 FormData

-0.06

े�

-0.06

iendo

-0.06

 eldest

-0.06

굴

-0.06

metatable

-0.06

ze

-0.06

Ske

-0.06

كيب

-0.06

POSITIVE LOGITS

 resolving

0.07

"display

0.07

 pretty

0.07

 canceled

0.07

 mascot

0.07

 range

0.06

 Fortune

0.06

 простран

0.06

—is

0.06

 ceramics

0.06

Activations Density 0.002%

weight loss

The neuron predominantly activates on self-referential, first-person words and constructions (e.g. “I,” “me,” “want,” “make”).

No Comments

No Known Activations