INDEX

Explanations

phrases related to names and locations

oai_token-act-pair · gpt-3.5-turbo

references to 'odge' and related names or terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

FIN

-0.74

 Steelers

-0.66

 Pirates

-0.66

dies

-0.66

 Penguins

-0.65

ername

-0.63

mpeg

-0.63

 Survivors

-0.63

 Humanity

-0.62

acious

-0.62

POSITIVE LOGITS

odge

1.06

 ethn

0.87

archy

0.85

chwitz

0.75

opol

0.75

 Lumpur

0.75

merce

0.75

isan

0.73

opping

0.73

cient

0.72

Activations Density 0.029%

phrases related to names and locations

references to 'odge' and related names or terms

No Comments

No Known Activations