INDEX

Explanations

phrases indicating continuity or connection

oai_token-act-pair · gpt-3.5-turbo

instances of the word "with" indicating companionship or association

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

ovich

-0.65

 Moody

-0.58

ynski

-0.58

undown

-0.56

atoes

-0.56

 excerpt

-0.55

NN

-0.55

heimer

-0.55

 midrange

-0.55

borgh

-0.55

POSITIVE LOGITS

stood

1.27

drawn

1.15

 impunity

1.01

draw

1.00

 regard

0.98

standing

0.97

holding

0.97

 regards

0.93

held

0.86

 dignity

0.86

Activations Density 0.158%

phrases indicating continuity or connection

instances of the word "with" indicating companionship or association

No Comments

No Known Activations