INDEX

Explanations

references to mathematical results and theorems within the context of scientific discussion

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

Not in Any Lists

No Comments

Negative Logits

 disambiguazione

-0.61

drawal

-0.59

EDEFAULT

-0.58

windowFixed

-0.56

 kasarigan

-0.55

tinyos

-0.54

Autoritní

-0.52

 للاسماء

-0.51

Демографія

-0.48

 Houſe

-0.48

POSITIVE LOGITS

новништво

0.45

 ظل

0.37

 okaza

0.37

 sledo

0.36

以上

0.34

シス

0.33

 ovan

0.33

 נוספת

0.33

freopen

0.32

ungszeit

0.31

Activations Density 0.115%

references to mathematical results and theorems within the context of scientific discussion

No Comments

No Known Activations