INDEX

Explanations

references to explosive devices

oai_token-act-pair · gpt-3.5-turbo

instances of the word "bomb" in various contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

laus

-0.87

SEA

-0.81

éĸ

-0.74

 Vale

-0.72

é»Ĵ

-0.70

 Princ

-0.68

Lic

-0.67

DragonMagazine

-0.67

 Legends

-0.66

 Dise

-0.66

POSITIVE LOGITS

arding

1.22

arded

1.12

 detonated

1.07

 bomb

1.00

 deton

1.00

 explodes

0.95

astically

0.95

 bombs

0.94

astic

0.91

blast

0.87

Activations Density 0.012%

references to explosive devices

instances of the word "bomb" in various contexts

No Comments

No Known Activations