INDEX

Explanations

references to military-related terms

oai_token-act-pair · gpt-3.5-turbo

references to military cadets

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Borders

-0.77

ter

-0.66

 luck

-0.65

 pity

-0.65

à

-0.64

 Mutual

-0.63

arenthood

-0.62

 Webster

-0.61

Ther

-0.61

ãĥīãĥ©

-0.60

POSITIVE LOGITS

cad

1.34

illac

1.00

 princ

1.00

Cad

0.88

encies

0.82

artisan

0.81

milo

0.80

 pige

0.80

illus

0.80

arbon

0.78

Activations Density 0.007%

references to military-related terms

references to military cadets

No Comments

No Known Activations