INDEX

Explanations

phrases indicating comparisons involving more than one or multiple entities

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

bufio

-0.73

gdx

-0.56

 respectivamente

-0.50

 berbeda

-0.48

expandindo

-0.46

 שונים

-0.45

styleable

-0.44

 inggris

-0.43

슷

-0.43

POSITIVE LOGITS

 usual

0.88

 ExecuteAsync

0.84

 StatelessWidget

0.83

kyou

0.83

 anything

0.81

 just

0.79

 THAN

0.77

>>>>>>>>

0.77

 ooit

0.77

 ever

0.76

Activations Density 0.107%

phrases indicating comparisons involving more than one or multiple entities

No Comments

No Known Activations