INDEX

Explanations

the word "Nam" at varying activation levels

oai_token-act-pair · gpt-3.5-turbo

references to specific names, particularly related to individuals and places

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

Introduced

-0.81

 Progressive

-0.74

Interview

-0.72

UID

-0.67

 Subtle

-0.65

ECB

-0.64

 subtitle

-0.62

 sample

-0.62

ACTED

-0.61

 Millenn

-0.61

POSITIVE LOGITS

nam

1.67

borgh

1.05

ukong

1.03

ned

1.03

orously

1.01

emi

0.95

emn

0.95

ovember

0.93

icol

0.93

rish

0.91

Activations Density 0.008%

the word "Nam" at varying activation levels

references to specific names, particularly related to individuals and places

No Comments

No Known Activations