INDEX

Explanations

patterns in coding structures and schema-related terms

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

Autoritní

-0.67

TagMode

-0.65

 snippetHide

-0.63

encodeWith

-0.60

 disambiguazione

-0.60

httphttps

-0.60

migrationBuilder

-0.58

 &___

-0.58

 ſind

-0.57

outSlope

-0.57

POSITIVE LOGITS

 around

0.35

 Around

0.33

 combattre

0.32

 largo

0.31

 exist

0.31

covery

0.30

тистика

0.30

 girar

0.30

 AROUND

0.29

CACHE

0.29

Activations Density 0.147%

patterns in coding structures and schema-related terms

No Comments

No Known Activations