INDEX

Explanations

mentions of violent or catastrophic events

oai_token-act-pair · gpt-3.5-turbo

instances of the word "bloody" and its variations in various contexts associated with violence or conflict

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

perm

-0.84

Reviewer

-0.79

berman

-0.77

Chart

-0.76

BOOK

-0.75

Introduced

-0.75

AUT

-0.74

Demand

-0.74

":[

-0.73

anol

-0.73

POSITIVE LOGITS

 bloody

1.12

 noses

0.81

 slaughter

0.81

 earthqu

0.80

 conflic

0.80

 Bloody

0.78

 prick

0.74

 wounds

0.72

 bruising

0.72

 teasp

0.70

Activations Density 0.005%

mentions of violent or catastrophic events

instances of the word "bloody" and its variations in various contexts associated with violence or conflict

No Comments

No Known Activations