INDEX

Explanations

the word "Till."

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and names related to events or contexts

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

 Magical

-0.79

usb

-0.69

hex

-0.69

des

-0.68

 broom

-0.67

occupied

-0.66

pots

-0.64

ZE

-0.63

UV

-0.63

che

-0.63

POSITIVE LOGITS

 Till

3.66

 Thur

1.43

Til

1.36

 Fuller

1.35

 Cull

1.34

 McCull

1.30

Ful

1.29

anth

1.17

eson

1.09

 Rudd

1.05

Activations Density 0.033%

the word "Till."

references to specific individuals and names related to events or contexts

No Comments

No Known Activations