INDEX

Explanations

elements related to testing and recommendations

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

caution or warning

np_acts-logits-general · gemini-2.0-flash

caution and advice

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 autorytatywna

-0.79

 unknownFields

-0.76

Portale

-0.73

 betweenstory

-0.66

 AssemblyCulture

-0.64

Personendaten

-0.59

HomeAsUp

-0.56

 Wikimedijinoj

-0.56

الإنجليزية

-0.55

quartered

-0.55

POSITIVE LOGITS

 beware

1.78

 Beware

1.66

Beware

1.62

 remember

1.59

注意

1.37

 Remember

1.32

Remember

1.31

 pastikan

1.31

 note

1.30

 caution

1.27

Activations Density 0.493%

elements related to testing and recommendations

caution or warning

caution and advice

No Comments

No Known Activations