INDEX

Explanations

elements related to testing and error reporting

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_21/width_16k/average_l0_70

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

Portale

-0.50

ьаж

-0.44

uche

-0.43

ضير

-0.42

 prim

-0.42

живать

-0.41

cives

-0.40

ICATED

-0.40

 middle

-0.39

pri

-0.39

POSITIVE LOGITS

 zero

1.91

 zéro

1.62

 Zero

1.59

zero

1.55

 ZERO

1.55

Zero

1.46

 cero

1.43

ZERO

1.31

 صفر

1.21

零

1.15

Activations Density 0.667%

elements related to testing and error reporting

No Comments

No Known Activations