INDEX

Explanations

references to medical data and patient statistics

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

medical patient outcomes

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

ſelves

-0.71

 pleaſure

-0.64

ſelf

-0.64

$_(

-0.63

$_"

-0.62

 ſche

-0.61

 verſ

-0.60

 iſt

-0.58

 Houſe

-0.57

 myſelf

-0.57

POSITIVE LOGITS

 commonest

0.49

__':

0.48

]=>

0.45

 retenir

0.44

libatkan

0.43

__":

0.40

出版年

0.40

 Millar

0.39

 hoog

0.38

 faltan

0.38

Activations Density 0.026%

references to medical data and patient statistics

medical patient outcomes

No Comments

No Known Activations