INDEX

Explanations

occurrences of the word "Autom" with varying levels of activation

oai_token-act-pair · gpt-3.5-turbo

references to "automation" and its related terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 McKenna

-0.71

 Michaels

-0.71

 Fitz

-0.68

 Hastings

-0.68

WHERE

-0.67

UGE

-0.67

EngineDebug

-0.66

patrick

-0.66

 straw

-0.66

 Byrne

-0.66

POSITIVE LOGITS

obiles

1.30

otive

1.28

obile

1.22

Autom

1.09

merce

1.01

 autom

0.95

aton

0.90

immune

0.88

atically

0.88

 Autom

0.87

Activations Density 0.005%

occurrences of the word "Autom" with varying levels of activation

references to "automation" and its related terms

No Comments

No Known Activations