INDEX

Explanations

phrases related to legislation and political discourse

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

stone

-0.67

bugs

-0.66

stones

-0.65

 âĢİ

-0.64

eteen

-0.63

eteenth

-0.62

oys

-0.62

idge

-0.61

IOR

-0.60

ocrin

-0.60

POSITIVE LOGITS

rew

0.78

 rightly

0.77

 preferably

0.75

 rightfully

0.70

 indeed

0.69

 perhaps

0.68

 hopefully

0.66

 therefore

0.66

 vice

0.64

 moreover

0.62

Activations Density 0.152%

phrases related to legislation and political discourse

No Comments

No Known Activations