INDEX

Explanations

phrases related to health and human services

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

"&" and "and"

np_acts-logits-general · gemini-2.0-flash

joining multiple terms

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

 rather

-0.70

شهاد

-0.62

 even

-0.61

 juist

-0.57

kwds

-0.55

windowFixed

-0.53

 also

-0.52

 both

-0.52

rån

-0.52

 asimismo

-0.50

POSITIVE LOGITS

 ostavi

0.71

]--;

0.64

Cyfeiriadau

0.64

LOYEES

0.62

ंदीखरीदारी

0.60

)_/¯

0.58

yaszt

0.57

amp

0.57

</tfoot>

0.57

crapers

0.56

Activations Density 0.208%

phrases related to health and human services

"&" and "and"

joining multiple terms

No Comments

No Known Activations