INDEX

Explanations

phrases related to organizations or company services

oai_token-act-pair · gpt-3.5-turbo

bullet points or list indicators in a text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

icum

-0.94

etheless

-0.91

iped

-0.84

alsa

-0.84

undai

-0.84

atur

-0.80

laus

-0.78

uly

-0.78

ajo

-0.77

ows

-0.77

POSITIVE LOGITS

âĢ¢âĢ¢

1.19

âĢ¢âĢ¢âĢ¢âĢ¢

1.09

ï¸ı

0.86

 FANTASY

0.81

âĸł

0.77

 âĢ¢

0.75

 DATA

0.73

Publisher

0.71

 Appears

0.69

Â½

0.69

Activations Density 0.019%

phrases related to organizations or company services

bullet points or list indicators in a text

No Comments

No Known Activations