INDEX

Explanations

the word "mean" followed by a digit, indicating an emphasis on interpretation or significance

oai_token-act-pair · gpt-3.5-turbo

phrases that convey the concept of meaning or significance

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

reb

-0.72

awaru

-0.61

 Ether

-0.59

~/.

-0.59

 Keeper

-0.59

 Reloaded

-0.57

 Emerging

-0.57

itiz

-0.57

 Scroll

-0.57

aea

-0.56

POSITIVE LOGITS

 mean

3.93

mean

2.76

 Mean

2.30

 means

1.61

 meant

1.48

 imply

1.39

 median

1.35

 signify

1.34

 Means

1.31

 entail

1.25

Activations Density 0.020%

the word "mean" followed by a digit, indicating an emphasis on interpretation or significance

phrases that convey the concept of meaning or significance

No Comments

No Known Activations