INDEX

Explanations

mentions of locations related to space such as NASA and the International Space Station

oai_token-act-pair · gpt-3.5-turbo

references to space-related topics

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

Negative Logits

kai

-0.75

tl

-0.74

ä»£

-0.73

æĢ

-0.68

UCT

-0.68

 presumptive

-0.66

venge

-0.65

inger

-0.64

CLOSE

-0.63

kt

-0.62

POSITIVE LOGITS

 Space

1.03

Space

1.03

 Spaces

0.94

 Horizons

0.87

bucks

0.81

 Orbit

0.80

 Shuttle

0.80

 Patrol

0.79

kat

0.78

 Astron

0.77

Activations Density 0.012%

mentions of locations related to space such as NASA and the International Space Station

references to space-related topics

No Known Activations