INDEX

Explanations

names of institutions and hospitals related to medical research

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

•korean

applies whole korean except first token of start

No Comments

Negative Logits

principalTable

-0.65

})->

-0.60

]")]

-0.57

 Alder

-0.55

 Plot

-0.54

 Horus

-0.54

 lenker

-0.53

signable

-0.53

ᅦ

-0.52

¬

-0.52

POSITIVE LOGITS

 Korean

1.18

 Korea

1.08

 Seoul

1.05

Korean

1.03

Korea

1.02

 korean

0.94

 Koreans

0.89

Seoul

0.89

 Corée

0.85

 coreano

0.83

Activations Density 0.170%

names of institutions and hospitals related to medical research

No Comments

No Known Activations