INDEX

Explanations

words related to empirical data and governmental terms such as 'Congress'

oai_token-act-pair · gpt-3.5-turbo

references to empirical research and legislative discussions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 2-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.2.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.2.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

zzle

-0.73

osity

-0.63

xual

-0.62

 lapse

-0.61

 Companies

-0.61

 payable

-0.60

 corporation

-0.58

 ceremony

-0.58

 grip

-0.57

ous

-0.57

POSITIVE LOGITS

ional

1.36

ions

1.09

icut

0.97

enegger

0.97

ãĥ¼ãĥĨãĤ£

0.96

ION

0.93

uates

0.93

uate

0.92

uating

0.91

uation

0.89

Activations Density 0.025%

words related to empirical data and governmental terms such as 'Congress'

references to empirical research and legislative discussions

No Comments

No Known Activations