INDEX

Explanations

phrases related to seriousness or severity

oai_token-act-pair · gpt-3.5-turbo

the term "serious" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

tein

-0.89

ucky

-0.76

ILA

-0.75

atu

-0.73

Pass

-0.72

AUT

-0.70

 dotted

-0.68

aver

-0.67

RIP

-0.66

eez

-0.65

POSITIVE LOGITS

serious

1.29

 serious

1.20

 seriousness

1.01

 Serious

0.97

 enough

0.80

geoning

0.78

lly

0.78

enough

0.77

 seriously

0.75

grave

0.74

Activations Density 0.012%

phrases related to seriousness or severity

the term "serious" in various contexts

No Comments

No Known Activations