INDEX

Explanations

backslashes followed by letters or numbers

oai_token-act-pair · gpt-3.5-turbo

references to formatting or organizational elements in text

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

;)

-0.84

:)

-0.76

[/

-0.74

:(

-0.73

!!!!

-0.73

.$

-0.72

âĢ¦âĢ¦âĢ¦âĢ¦

-0.71

!!!

-0.69

!.

-0.69

 ðŁĻĤ

-0.69

POSITIVE LOGITS

quartered

0.47

heny

0.47

ãĥ«

0.46

worm

0.45

espie

0.45

jamin

0.42

owing

0.41

 Americans

0.41

 American

0.41

beans

0.40

Activations Density 1.863%

backslashes followed by letters or numbers

references to formatting or organizational elements in text

No Comments

No Known Activations