INDEX

Explanations

photographs in the text

oai_token-act-pair · gpt-3.5-turbo

references to photographs

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

senal

-0.93

uggest

-0.93

 stocking

-0.90

ingers

-0.90

ighters

-0.88

pace

-0.79

vik

-0.77

pring

-0.76

gaard

-0.76

 eager

-0.75

POSITIVE LOGITS

 photograph

1.04

 PHOTO

0.89

Thumbnail

0.89

SHARE

0.87

 Photograph

0.83

 Flickr

0.83

enance

0.82

 Shoot

0.78

 photo

0.78

ogram

0.78

Activations Density 0.015%

photographs in the text

references to photographs

No Comments

No Known Activations