INDEX

Explanations

terms related to descriptions or classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 fédé

-0.41

 fédéral

-0.39

 juges

-0.36

 précédentes

-0.34

 Verständnis

-0.34

 însă

-0.34

 plazos

-0.34

 quizá

-0.34

 juridiques

-0.34

 précédente

-0.33

POSITIVE LOGITS

 called

0.90

 CALLED

0.87

called

0.85

NameInMap

0.85

VersionUID

0.79

 Called

0.79

Personendaten

0.77

Called

0.76

 termed

0.74

称为

0.74

Activations Density 0.400%

terms related to descriptions or classifications

No Comments

No Known Activations