INDEX

Explanations

locations or structures typically found in public spaces

oai_token-act-pair · gpt-3.5-turbo

references to kiosks and public spaces

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

mitted

-0.64

 Mend

-0.63

 Sussex

-0.63

 Sabbath

-0.61

 uncomp

-0.61

IPM

-0.61

Dad

-0.61

Ð´

-0.60

 Sabb

-0.60

 Parents

-0.60

POSITIVE LOGITS

 kios

1.03

 restrooms

1.01

cers

0.97

 restroom

0.92

 seating

0.92

itory

0.92

staff

0.90

ĸļ

0.83

 booths

0.82

rette

0.81

Activations Density 0.029%

locations or structures typically found in public spaces

references to kiosks and public spaces

No Comments

No Known Activations