INDEX

Explanations

words related to the design or planning of products or systems

oai_token-act-pair · gpt-3.5-turbo

instances of the word "designed" in various contexts

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

bour

-0.60

some

-0.60

 Michaels

-0.59

 Ended

-0.58

owners

-0.58

asta

-0.58

Zone

-0.57

 Sussex

-0.57

chal

-0.56

rika

-0.55

POSITIVE LOGITS

 specifically

1.00

 expressly

0.95

to

0.83

 specially

0.80

irection

0.79

 designed

0.78

 accordingly

0.78

 primarily

0.78

 solely

0.74

designed

0.74

Activations Density 0.068%

words related to the design or planning of products or systems

instances of the word "designed" in various contexts

No Comments

No Known Activations