INDEX

Explanations

keywords related to cups or containers

oai_token-act-pair · gpt-3.5-turbo

references to "cup" in various contexts

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

IGHTS

-0.69

doms

-0.66

 Giuliani

-0.65

ccording

-0.64

eering

-0.63

OUND

-0.62

 gearing

-0.62

 antid

-0.62

 Kazakh

-0.61

Voc

-0.61

POSITIVE LOGITS

cakes

1.74

cake

1.71

cup

1.21

board

0.95

 cups

0.91

maid

0.91

cup

0.89

boards

0.87

illo

0.86

rant

0.82

Activations Density 0.013%

keywords related to cups or containers

references to "cup" in various contexts

No Comments

No Known Activations