INDEX

Explanations

references to the superhero character "Wonder Woman."

oai_token-act-pair · gpt-3.5-turbo

references to "Wonder Woman"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

escription

-0.75

 externalToEVAOnly

-0.73

 destro

-0.71

aution

-0.71

 seiz

-0.65

ABE

-0.65

 manslaughter

-0.63

 VIDE

-0.63

Obj

-0.63

 disg

-0.61

POSITIVE LOGITS

kid

1.12

stru

1.12

fully

1.09

bolt

1.07

ful

1.06

FUL

1.05

kids

1.03

fulness

1.01

wall

0.98

bol

0.98

Activations Density 0.037%

references to the superhero character "Wonder Woman."

references to "Wonder Woman"

No Comments

No Known Activations