INDEX

Explanations

phrases related to design and intention in various contexts

oai_token-act-pair · gpt-3.5-turbo

instances of the word "designed."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 4-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.4.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.4.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

dash

-0.73

 Ezek

-0.68

enne

-0.68

nikov

-0.66

held

-0.65

Isa

-0.64

Charge

-0.63

rika

-0.63

 Michaels

-0.63

hus

-0.63

POSITIVE LOGITS

urally

0.98

ators

0.82

 flaw

0.76

 Parenthood

0.76

ural

0.75

 layouts

0.74

ating

0.72

yout

0.71

 designing

0.70

ĸļ

0.70

Activations Density 0.029%

phrases related to design and intention in various contexts

instances of the word "designed."

No Comments

No Known Activations