INDEX

Explanations

selfies, both being taken and posted

oai_token-act-pair · gpt-3.5-turbo

references to selfies 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

endez

-0.84

profits

-0.76

endale

-0.76

ONSORED

-0.75

ends

-0.74

iance

-0.74

 Flavoring

-0.73

end

-0.73

iant

-0.73

artment

-0.72

POSITIVE LOGITS

 selfies

1.07

 selfie

0.99

OPLE

0.75

 Gallery

0.75

 Doodle

0.73

 onstage

0.73

tan

0.70

 pics

0.69

 nude

0.69

 outdoors

0.68

Activations Density 0.026%

selfies, both being taken and posted

references to selfies and related actions

No Comments

No Known Activations