INDEX

Explanations

phrases related to various scenarios, including experiences with certain conditions and discussing topics in a nuanced or precise way

oai_token-act-pair · gpt-3.5-turbo

instances of the word "have."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 3-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.3.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.3.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

 Apart

-0.75

catentry

-0.64

icking

-0.58

inance

-0.56

oshi

-0.56

 fiasco

-0.55

arter

-0.55

 debacle

-0.54

osa

-0.54

 settlement

-0.54

POSITIVE LOGITS

 been

1.36

been

1.15

 Been

1.06

 gotten

1.02

 seen

0.93

 undergone

0.92

 gone

0.88

gotten

0.88

 begun

0.87

kell

0.87

Activations Density 0.316%

phrases related to various scenarios, including experiences with certain conditions and discussing topics in a nuanced or precise way

instances of the word "have."

No Comments

No Known Activations