INDEX

Explanations

references to the TV show "WWE Monday Night Raw."

oai_token-act-pair · gpt-3.5-turbo

references to the WWE show "Raw."

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

etheless

-0.84

uate

-0.84

repre

-0.80

uated

-0.79

ortium

-0.73

anwhile

-0.71

raints

-0.70

mercial

-0.70

arios

-0.69

 therap

-0.69

POSITIVE LOGITS

lings

1.32

lins

1.20

hide

1.19

ls

1.17

leigh

0.98

kus

0.90

esome

0.88

burn

0.86

cliffe

0.86

ford

0.85

Activations Density 0.004%

references to the TV show "WWE Monday Night Raw."

references to the WWE show "Raw."

No Comments

No Known Activations