INDEX

Explanations

phrases related to law enforcement, protests, and armed conflicts

oai_token-act-pair · gpt-3.5-turbo

references to police and military actions, particularly those involving crowd control and weaponry

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

erest

-0.82

izens

-0.78

Disney

-0.76

ittal

-0.74

iverse

-0.72

*/(

-0.70

arenthood

-0.70

iquid

-0.70

(<

-0.70

izen

-0.69

POSITIVE LOGITS

 grenades

1.63

 rifles

1.50

 explosives

1.50

 shotguns

1.46

 knives

1.44

 pistols

1.43

 gunfire

1.43

 torches

1.42

 snipers

1.37

 guns

1.36

Activations Density 0.315%

phrases related to law enforcement, protests, and armed conflicts

references to police and military actions, particularly those involving crowd control and weaponry

No Comments

No Known Activations