INDEX

Explanations

words and phrases related to searching for solutions or information

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

"can" or its translations

np_acts-logits-general · gemini-2.0-flash

ability or possibility

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

rawDesc

-0.44

interopRequire

-0.42

pides

-0.41

Were

-0.41

zielle

-0.40

 INTERESAR

-0.40

 yoksa

-0.40

tvrt

-0.40

 Were

-0.39

Населення

-0.39

POSITIVE LOGITS

can

3.00

can

2.31

 could

2.12

Can

2.09

 bisa

2.05

Can

1.98

 dapat

1.96

可以

1.95

 können

1.92

 можно

1.85

Activations Density 2.274%

words and phrases related to searching for solutions or information

"can" or its translations

ability or possibility

No Comments

No Known Activations