INDEX

Explanations

the acronym "AN" followed by a single capitalized word

oai_token-act-pair · gpt-3.5-turbo

references to the abbreviation "AN" in various contexts

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

lahoma

-0.78

ãĤ¡

-0.73

 guiIcon

-0.72

 Wink

-0.70

 Jarrett

-0.68

åĭ

-0.67

 Gemini

-0.67

Kem

-0.67

ï¸ı

-0.65

odore

-0.65

POSITIVE LOGITS

igans

1.23

AN

1.13

ARCH

0.98

ufact

0.98

IAN

0.92

ESE

0.92

thood

0.92

ICAN

0.90

alyst

0.88

CAR

0.88

Activations Density 0.008%

the acronym "AN" followed by a single capitalized word

references to the abbreviation "AN" in various contexts

No Comments

No Known Activations