INDEX

Explanations

specific terms related to leadership roles and bureaucratic functions

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

 Park

-0.41

 park

-0.40

CP

-0.39

Bl

-0.38

 Master

-0.37

 blind

-0.37

Véxase

-0.36

 оригіналу

-0.35

POSITIVE LOGITS

 gynhyrchwyd

0.60

Попис

0.57

 dévelo

0.56

 yoksa

0.55

ftagPool

0.55

 exportaciones

0.55

harapkan

0.54

 dentaire

0.53

 defaultstate

0.53

 الرياضيه

0.52

Activations Density 0.099%

specific terms related to leadership roles and bureaucratic functions

No Comments

No Known Activations