INDEX

Explanations

words related to prototyping and testing

oai_token-act-pair · gpt-3.5-turbo

terms related to prototyping and stereotypes

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

waters

-0.73

wife

-0.70

 Olsen

-0.66

IRO

-0.62

 thereto

-0.60

 ACTION

-0.59

WD

-0.59

 jersey

-0.57

Rae

-0.57

 prisoner

-0.56

POSITIVE LOGITS

otyp

1.28

ically

1.06

itives

1.05

ing

1.03

egu

1.01

inar

0.99

ifact

0.99

ed

0.98

icol

0.97

icative

0.97

Activations Density 0.024%

words related to prototyping and testing

terms related to prototyping and stereotypes

No Comments

No Known Activations