INDEX

Explanations

words related to various communication protocols or standards

oai_token-act-pair · gpt-3.5-turbo

specific nouns or proper names, often related to notable entities or concepts

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

acebook

-1.19

 referen

-1.09

efeated

-1.07

vertisement

-1.02

retty

-1.01

ometimes

-0.99

GGGGGGGG

-0.95

razil

-0.92

mble

-0.91

iflower

-0.89

POSITIVE LOGITS

 Lumpur

1.03

dain

0.77

×Ĳ

0.77

×ª

0.75

lust

0.71

×ŀ

0.71

BILITY

0.70

×

0.67

WAYS

0.67

illion

0.67

Activations Density 1.404%

words related to various communication protocols or standards

specific nouns or proper names, often related to notable entities or concepts

No Comments

No Known Activations