INDEX

Explanations

phrases indicating lack of authorization or permission

oai_token-act-pair · gpt-3.5-turbo

phrases related to authority and responsibility within a corporate context

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

ops

-0.69

 aloud

-0.67

 PsyNet

-0.66

oggle

-0.66

culosis

-0.66

ateurs

-0.65

reprene

-0.64

*/(

-0.64

enz

-0.64

atives

-0.63

POSITIVE LOGITS

 dimensional

0.68

ãĥ©ãĥ³

0.67

pal

0.60

 Sparrow

0.60

thy

0.59

 stabilized

0.59

hemer

0.58

 collateral

0.58

 majesty

0.57

 Sabbath

0.57

Activations Density 0.000%

phrases indicating lack of authorization or permission

phrases related to authority and responsibility within a corporate context

No Comments

No Known Activations