INDEX

Explanations

proper nouns, possibly referring to political figures

oai_token-act-pair · gpt-3.5-turbo

references to "Prime Minister."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

 sheet

-0.69

 wing

-0.67

 pill

-0.67

 Atlanta

-0.66

hub

-0.64

LA

-0.64

 Florida

-0.63

 field

-0.63

 discipline

-0.63

 drugs

-0.62

POSITIVE LOGITS

Prime

4.02

prime

1.95

 Prime

1.65

Prem

1.36

 prime

1.30

Mayor

1.27

Prim

1.23

Pope

1.22

Labour

1.20

President

1.19

Activations Density 0.020%

proper nouns, possibly referring to political figures

references to "Prime Minister."

No Comments

No Known Activations