INDEX

Explanations

words related to opinions or actions being unpopular

oai_token-act-pair · gpt-3.5-turbo

references to unpopularity

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

EStreamFrame

-0.81

erning

-0.78

ramid

-0.76

initely

-0.75

hens

-0.75

chn

-0.74

llular

-0.72

utics

-0.71

arnaev

-0.71

ynthesis

-0.70

POSITIVE LOGITS

ity

1.19

 unpopular

1.12

 incumbent

0.94

ities

0.89

 majorities

0.82

 burdens

0.76

 incumb

0.75

nesses

0.74

lihood

0.73

 taboo

0.72

Activations Density 0.021%

words related to opinions or actions being unpopular

references to unpopularity

No Comments

No Known Activations