INDEX

Explanations

Code/online content

np_max-act · gemini-2.0-flash

Tokens marking the assistant speaker/assistant-message header (i.e., the "assistant" role token in the chat metadata).

oai_token-act-pair · gpt-5-mini Triggered by @vetterc0

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_15/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.15.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 longitud

-0.07

esture

-0.06

Sys

-0.06

 Thanh

-0.06

is

-0.06

�

-0.06

(high

-0.06

Harry

-0.06

 Ром

-0.06

 além

-0.06

POSITIVE LOGITS

的な

0.07

 inaccessible

0.07

Yap

0.07

ていない

0.06

 nationalist

0.06

 advertisement

0.06

_new

0.06

 werd

0.06

 "-"↵

0.06

 Boise

0.06

Activations Density 0.109%

Code/online content

Tokens marking the assistant speaker/assistant-message header (i.e., the "assistant" role token in the chat metadata).

No Comments

No Known Activations