INDEX

Explanations

keywords related to programming or coding

oai_token-act-pair · gpt-3.5-turbo

the prefix "Pre" at the beginning of words

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

 tears

-0.75

 wonder

-0.72

 entertain

-0.66

odd

-0.65

 dynam

-0.64

 darts

-0.64

 infinity

-0.63

 elevator

-0.63

Yon

-0.63

 laughter

-0.62

POSITIVE LOGITS

Pre

3.55

pre

2.16

PRE

2.03

Pre

2.00

PRE

1.68

Prep

1.57

Pref

1.52

pre

1.43

Prior

1.36

 Prep

1.28

Activations Density 0.015%

keywords related to programming or coding

the prefix "Pre" at the beginning of words

No Comments

No Known Activations