INDEX

Explanations

photographs or images

oai_token-act-pair · gpt-3.5-turbo

references to photographs or images in the text

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

worldly

-0.76

astic

-0.73

astically

-0.72

soc

-0.70

asons

-0.69

ousy

-0.65

cients

-0.64

ressive

-0.64

iatric

-0.64

tower

-0.64

POSITIVE LOGITS

 Photographer

0.77

 IMAGES

0.72

HAEL

0.71

 caption

0.70

SHARE

0.70

 guiActiveUnfocused

0.68

 Phot

0.65

 Vehicles

0.65

 Uriel

0.65

 Aerial

0.64

Activations Density 0.036%

photographs or images

references to photographs or images in the text

No Comments

No Known Activations