INDEX

Explanations

keywords related to clicks or clicking on links

oai_token-act-pair · gpt-3.5-turbo

phrases related to clicking and engagement with links or advertisements

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

egal

-0.68

 Variety

-0.67

 Yard

-0.67

ZA

-0.63

 Chancellor

-0.62

 Scotia

-0.62

 Ministers

-0.61

 Arabs

-0.60

 Catalonia

-0.58

 Territories

-0.58

POSITIVE LOGITS

lish

0.93

prints

0.87

dress

0.84

through

0.83

pad

0.82

ety

0.80

esome

0.79

ername

0.77

antry

0.75

wheel

0.73

Activations Density 0.035%

keywords related to clicks or clicking on links

phrases related to clicking and engagement with links or advertisements

No Comments

No Known Activations