INDEX

Explanations

phrases related to change and progression

oai_token-act-pair · gpt-3.5-turbo

instances of the word "and" indicating connections or continuations in thought

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

asar

-0.74

uta

-0.73

ut

-0.72

ident

-0.72

GY

-0.70

shift

-0.68

icts

-0.68

acas

-0.67

])

-0.67

agon

-0.66

POSITIVE LOGITS

 furthermore

1.07

 moreover

1.05

 rightfully

1.03

 therefore

1.00

 consequently

0.96

ersen

0.94

 rightly

0.92

 therein

0.91

 congr

0.91

 secondly

0.90

Activations Density 0.231%

phrases related to change and progression

instances of the word "and" indicating connections or continuations in thought

No Comments

No Known Activations