INDEX

Explanations

professions, expertise, and areas of importance

oai_token-act-pair · gpt-3.5-turbo

phrases indicating significance or interest

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

 Cobb

-0.69

 Wraith

-0.61

 Doomsday

-0.59

 caval

-0.59

RANT

-0.59

awa

-0.57

 Camden

-0.57

 Viper

-0.56

BMC

-0.56

Raw

-0.55

POSITIVE LOGITS

 proportions

0.78

 fame

0.71

mods

0.69

chuk

0.67

yout

0.67

interest

0.66

lain

0.65

 persuasion

0.64

farious

0.64

interstitial

0.63

Activations Density 0.257%

professions, expertise, and areas of importance

phrases indicating significance or interest

No Comments

No Known Activations