INDEX

Explanations

words related to action figures and collectible figurines

oai_token-act-pair · gpt-3.5-turbo

references to action figures and collectibles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

lished

-0.93

rup

-0.83

artney

-0.81

bern

-0.78

estern

-0.77

NPR

-0.75

velength

-0.74

bey

-0.74

termination

-0.73

upuncture

-0.72

POSITIVE LOGITS

 figur

0.96

 wraps

0.87

 sculptures

0.81

 figures

0.74

 figure

0.74

 sculpture

0.74

 sculpt

0.74

 statues

0.73

 wrap

0.72

 dolls

0.72

Activations Density 0.030%

words related to action figures and collectible figurines

references to action figures and collectibles

No Comments

No Known Activations