INDEX

Explanations

parentheses indicating numerical values with high importance

oai_token-act-pair · gpt-3.5-turbo

parentheses or bracketed content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Academy

-0.71

 glare

-0.71

 Lumpur

-0.70

zoo

-0.68

 Lynd

-0.68

 wildlife

-0.66

 resur

-0.66

sav

-0.66

 park

-0.65

 lull

-0.64

POSITIVE LOGITS

including

1.65

excluding

1.54

such

1.51

typically

1.48

which

1.44

usually

1.42

often

1.39

except

1.38

either

1.38

meaning

1.37

Activations Density 0.153%

parentheses indicating numerical values with high importance

parentheses or bracketed content

No Comments

No Known Activations