INDEX

Explanations

phrases related to emphasizing or pointing out specific words or phrases

oai_token-act-pair · gpt-3.5-turbo

references to articles and their content

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

ordable

-0.83

habi

-0.83

ornings

-0.79

urnal

-0.78

adle

-0.76

elaide

-0.76

soever

-0.76

leground

-0.75

thood

-0.73

entimes

-0.69

POSITIVE LOGITS

 implication

1.57

 analogy

1.48

 gist

1.45

 wording

1.36

 distinction

1.33

 argument

1.30

 assumption

1.27

 reasoning

1.25

 inference

1.24

 difference

1.22

Activations Density 0.523%

phrases related to emphasizing or pointing out specific words or phrases

references to articles and their content

No Comments

No Known Activations