INDEX

Explanations

terms related to communication and processes within systems

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

 plegable

-0.49

 touches

-0.47

 initComponents

-0.45

idać

-0.43

PLUG

-0.43

 skipped

-0.42

 çocuklar

-0.41

komsten

-0.41

 Touch

-0.40

 TOUCH

-0.40

POSITIVE LOGITS

RTDA

0.53

tagHelperRunner

0.52

CppMethod

0.50

iNdEx

0.50

ImageContext

0.48

 PyTuple

0.46

 onOptions

0.44

嘩

0.43

zzleHttp

0.42

wpdb

0.42

Activations Density 0.056%

terms related to communication and processes within systems

No Comments

No Known Activations