INDEX

Explanations

references to derogatory or critical expressions

oai_token-act-pair · gpt-3.5-turbo

words related to humor and satire

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

rompt

-0.66

ioxide

-0.65

 luck

-0.64

 Helpful

-0.64

 concess

-0.64

resa

-0.63

 sincerity

-0.63

ãĥ¼ãĤ¯

-0.63

 constitu

-0.62

 Archdemon

-0.62

POSITIVE LOGITS

auga

0.78

ards

0.73

pole

0.72

rake

0.72

dden

0.72

eston

0.71

ills

0.69

aceous

0.69

hess

0.69

ppings

0.69

Activations Density 0.141%

references to derogatory or critical expressions

words related to humor and satire

No Comments

No Known Activations