INDEX

Explanations

references to walking and walks

oai_token-act-pair · gpt-3.5-turbo

references to walking activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

cffff

-0.80

uchin

-0.71

ccording

-0.71

illian

-0.71

uliffe

-0.71

milo

-0.71

ultz

-0.69

ARA

-0.65

ancies

-0.64

berra

-0.64

POSITIVE LOGITS

Walk

1.01

walk

0.95

 Walk

0.94

 walk

0.91

ways

0.89

walker

0.87

athon

0.87

own

0.86

lihood

0.84

walking

0.83

Activations Density 0.017%

references to walking and walks

references to walking activities

No Comments

No Known Activations