INDEX

Explanations

references to legal or criminal activities, investigations, and accusations

oai_token-act-pair · gpt-3.5-turbo

phrases related to controversial or negative outcomes

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

itialized

-0.64

OUP

-0.59

eport

-0.58

onom

-0.55

rophe

-0.55

 Powered

-0.53

iple

-0.53

GW

-0.52

uum

-0.49

 Redmond

-0.49

POSITIVE LOGITS

 thereby

0.78

anwhile

0.67

 respectively

0.66

0.63

)).

0.62

SPONSORED

0.59

consider

0.59

 consequently

0.59

thus

0.59

 culminating

0.58

Activations Density 1.688%

references to legal or criminal activities, investigations, and accusations

phrases related to controversial or negative outcomes

No Comments

No Known Activations