INDEX

Explanations

words related to male names or entities containing "Wh"

oai_token-act-pair · gpt-3.5-turbo

repeated instances of the character sequence "Wh"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

CVE

-0.92

actionDate

-0.82

alian

-0.79

coded

-0.73

 appendix

-0.73

CTR

-0.66

fecture

-0.66

Ð°

-0.63

hap

-0.62

ANI

-0.61

POSITIVE LOGITS

Wh

3.70

Wh

2.49

wh

1.97

wh

1.76

WH

1.65

 Whale

1.52

Whe

1.43

 Whis

1.39

WH

1.29

 Whit

1.26

Activations Density 0.013%

words related to male names or entities containing "Wh"

repeated instances of the character sequence "Wh"

No Comments

No Known Activations