INDEX

Explanations

information related to news segments and interviews

oai_token-act-pair · gpt-3.5-turbo

expressions related to uncertainty or speculation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

 opting

-0.68

 grooming

-0.67

loo

-0.65

 styl

-0.65

 Mane

-0.65

 favourable

-0.64

 snowball

-0.64

 Machina

-0.64

 composing

-0.64

 pursuing

-0.63

POSITIVE LOGITS

BILL

1.39

AMY

1.37

 PRESIDENT

1.28

REP

1.23

AUD

1.18

PRES

1.16

TRUMP

1.16

QUEST

1.15

STEP

1.14

JOHN

1.10

Activations Density 0.172%

information related to news segments and interviews

expressions related to uncertainty or speculation

No Comments

No Known Activations