INDEX

Explanations

phrases related to negation

oai_token-act-pair · gpt-3.5-turbo

negative prefixes attached to adjectives or nouns

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

ħĭ

-0.91

ulhu

-0.88

 Lumpur

-0.77

 Dickinson

-0.76

Zup

-0.76

 Cassidy

-0.75

ĸļ

-0.71

wagen

-0.69

 Brus

-0.68

essee

-0.67

POSITIVE LOGITS

existent

1.27

profit

1.05

conscious

1.00

zero

0.99

issue

0.98

exclusive

0.98

human

0.97

issues

0.96

responsive

0.95

cont

0.94

Activations Density 0.034%

phrases related to negation

negative prefixes attached to adjectives or nouns

No Comments

No Known Activations