INDEX

Explanations

the word "Gaz" at various activations

oai_token-act-pair · gpt-3.5-turbo

mentions of specific geographical locations and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

ACTED

-0.78

enment

-0.73

 Fargo

-0.70

icio

-0.66

EED

-0.64

velength

-0.64

aceous

-0.63

 Wooden

-0.62

 Crane

-0.61

 Hawaiian

-0.60

POSITIVE LOGITS

hou

0.98

bour

0.97

rina

0.94

prom

0.91

illion

0.88

low

0.87

mos

0.85

den

0.83

busters

0.83

akov

0.83

Activations Density 0.043%

the word "Gaz" at various activations

mentions of specific geographical locations and organizations

No Comments

No Known Activations