INDEX

Explanations

phrases indicating that something has already been done or decided previously

oai_token-act-pair · gpt-3.5-turbo

the phrase "you've" and its variations

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eem

-0.64

 Accounts

-0.64

WARD

-0.62

ace

-0.62

 manifests

-0.61

soever

-0.61

 Torrent

-0.60

 denotes

-0.59

 signifies

-0.58

 Lich

-0.57

POSITIVE LOGITS

 gotten

1.46

 been

1.45

got

1.44

 gotta

1.41

been

1.28

 Been

1.19

 done

1.15

 seen

1.13

 gone

1.11

gotten

1.09

Activations Density 0.073%

phrases indicating that something has already been done or decided previously

the phrase "you've" and its variations

No Comments

No Known Activations