INDEX

Explanations

mentions of romantic or intimate actions like kissing

oai_token-act-pair · gpt-3.5-turbo

references to kissing and related actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

é¾

-1.01

ifted

-0.74

quickShipAvailable

-0.73

ulhu

-0.70

izoph

-0.70

æ©Ł

-0.67

IJ

-0.67

 Administ

-0.67

ENCY

-0.67

 DISTR

-0.66

POSITIVE LOGITS

 goodbye

1.28

 kiss

1.14

 Kiss

1.07

 kisses

1.06

 kissing

0.99

kiss

0.99

creen

0.98

 kissed

0.95

 passionately

0.91

 Goodbye

0.83

Activations Density 0.026%

mentions of romantic or intimate actions like kissing

references to kissing and related actions

No Comments

No Known Activations