INDEX

Explanations

references to news agencies like AP in the text

oai_token-act-pair · gpt-3.5-turbo

instances of the Associated Press (AP) news source

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 stabil

-0.70

ãĤ©

-0.69

 thumbs

-0.69

 capitalist

-0.66

 gluten

-0.64

ãĥĥãĥī

-0.64

 ment

-0.64

 phyl

-0.63

Ca

-0.63

 Kinn

-0.61

POSITIVE LOGITS

PLIC

1.17

PLIED

1.16

PLE

1.06

PROV

0.99

TN

0.95

ocalypse

0.95

oleon

0.94

rison

0.93

OE

0.93

ropri

0.92

Activations Density 0.009%

references to news agencies like AP in the text

instances of the Associated Press (AP) news source

No Comments

No Known Activations