INDEX

Explanations

mathematical relationships or comparisons with the word "equal"

oai_token-act-pair · gpt-3.5-turbo

equivalences and comparisons related to values or quantities

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

ãĤ©

-0.72

offer

-0.70

 Franch

-0.66

wig

-0.66

fried

-0.65

å£

-0.65

nc

-0.62

holiday

-0.62

 Fenrir

-0.62

 Kitt

-0.61

POSITIVE LOGITS

 equals

3.26

imble

1.06

 onset

0.97

kin

0.93

 Loading

0.90

ELY

0.86

 containment

0.85

ilty

0.85

usat

0.84

 constructor

0.78

Activations Density 0.058%

mathematical relationships or comparisons with the word "equal"

equivalences and comparisons related to values or quantities

No Comments

No Known Activations