INDEX

Explanations

specialized terminology or jargon related to technology and media

oai_token-act-pair · gpt-3.5-turbo

references to potential threats and complex situations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

âĢ

-0.95

à¨

-0.94

âĸĳ

-0.81

Â¯Â¯

-0.81

à©

-0.78

ntil

-0.77

few

-0.75

 à¨

-0.75

âĹ

-0.75

ãĥĥãĥī

-0.75

POSITIVE LOGITS

!:

0.72

0.68

!".

0.64

to

0.64

!'

0.62

TO

0.62

To

0.61

!'"

0.60

!]

0.59

 Spectre

0.59

Activations Density 1.124%

specialized terminology or jargon related to technology and media

references to potential threats and complex situations

No Comments

No Known Activations