INDEX

Explanations

adjectives and nouns related to negative or disturbing events or characteristics

oai_token-act-pair · gpt-3.5-turbo

terms related to atrocities and decadence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 7-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.7.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.7.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

ixel

-0.79

quickShipAvailable

-0.79

arten

-0.78

glas

-0.75

ocular

-0.73

 Boards

-0.72

RFC

-0.68

 GOODMAN

-0.67

yip

-0.66

soType

-0.66

POSITIVE LOGITS

 inflicted

1.18

 perpetrated

0.99

 havoc

0.97

 scourge

0.95

 imaginable

0.88

 ridden

0.87

 spew

0.87

 plag

0.87

 Incarn

0.87

wre

0.86

Activations Density 0.297%

adjectives and nouns related to negative or disturbing events or characteristics

terms related to atrocities and decadence

No Comments

No Known Activations