INDEX

Explanations

parentheses and technical information within text

oai_token-act-pair · gpt-3.5-turbo

parentheses and brackets that denote references or citations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 coales

-0.74

 conver

-0.64

 clin

-0.64

 squared

-0.61

 outl

-0.60

 langu

-0.59

 elevated

-0.58

 assigned

-0.58

 expectancy

-0.57

 pronouns

-0.56

POSITIVE LOGITS

AFP

1.43

VIDEO

1.41

Picture

1.39

PHOTOS

1.37

Photo

1.37

photo

1.28

Reuters

1.25

Photos

1.20

Adds

1.20

CBC

1.18

Activations Density 0.159%

parentheses and technical information within text

parentheses and brackets that denote references or citations

No Comments

No Known Activations