INDEX

Explanations

the word "Grant"

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Grant"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

dayName

-0.79

computer

-0.71

 abdom

-0.67

-0.66

ä¸ī

-0.65

ãĥķãĤ©

-0.64

ptroller

-0.64

Assembly

-0.63

Temperature

-0.62

Downloadha

-0.62

POSITIVE LOGITS

 Grant

1.00

Grant

0.94

lee

0.92

ees

0.88

 Morrison

0.88

anas

0.86

 Thornton

0.85

aire

0.81

elson

0.77

witz

0.76

Activations Density 0.007%

the word "Grant"

mentions of the name "Grant"

No Comments

No Known Activations