INDEX

Explanations

mentions of a specific term, "cancer-causing."

oai_token-act-pair · gpt-3.5-turbo

references to a specific type of cancer or carcinogenic substances

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

pocket

-0.80

schild

-0.75

UID

-0.75

 GOODMAN

-0.73

 outfield

-0.66

hower

-0.66

enegger

-0.65

geist

-0.65

externalActionCode

-0.64

states

-0.64

POSITIVE LOGITS

ption

1.12

ca

1.01

ffe

0.98

esar

0.92

utical

0.91

xon

0.90

usa

0.90

ensis

0.85

0.84

fu

0.83

Activations Density 0.006%

mentions of a specific term, "cancer-causing."

references to a specific type of cancer or carcinogenic substances

No Comments

No Known Activations