INDEX

Explanations

the word 'let'

oai_token-act-pair · gpt-3.5-turbo Triggered by @bot

instances of the word "let" indicating permission or allowance

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 12-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.hook_resid_post

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.11.hook_resid_post

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

 Recent

-0.60

 Congratulations

-0.59

esi

-0.57

ombat

-0.56

 Instruct

-0.55

inger

-0.55

assian

-0.54

cill

-0.53

lihood

-0.52

acci

-0.52

POSITIVE LOGITS

 alone

1.57

tered

1.25

tering

1.05

downs

0.96

ting

0.93

icia

0.89

 loose

0.88

down

0.85

ogether

0.83

ãĥĥãĥī

0.77

Activations Density 0.019%

the word 'let'

instances of the word "let" indicating permission or allowance

No Comments

No Known Activations