INDEX

Explanations

mentions of the name "Andy" at various activation strengths

oai_token-act-pair · gpt-3.5-turbo

mentions of the name "Andy" in the text

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

atern

-0.80

orescent

-0.78

regulated

-0.78

racted

-0.73

restricted

-0.69

ĺħ

-0.69

 affili

-0.69

 nationally

-0.69

ornia

-0.68

 confidential

-0.67

POSITIVE LOGITS

andy

1.12

wine

0.98

 Candy

0.88

oga

0.84

 Melody

0.84

rand

0.83

 Slug

0.77

 Ples

0.77

 Crush

0.76

 Crowley

0.76

Activations Density 0.003%

mentions of the name "Andy" at various activation strengths

mentions of the name "Andy" in the text

No Comments

No Known Activations