INDEX

Explanations

"stylishly" presented information or descriptions

oai_token-act-pair · gpt-3.5-turbo

words related to style and self-identification

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

 Americas

-0.73

 Dickinson

-0.70

 outweigh

-0.68

 punt

-0.66

 microsc

-0.66

 exits

-0.65

 Rush

-0.64

 Wire

-0.64

kin

-0.64

 intra

-0.64

POSITIVE LOGITS

appointed

2.83

sty

2.75

proclaimed

2.65

described

2.42

identified

1.60

decl

1.55

imposed

1.31

prof

1.29

named

1.29

employed

1.19

Activations Density 0.041%

"stylishly" presented information or descriptions

words related to style and self-identification

No Comments

No Known Activations