INDEX

Explanations

proper nouns, specifically names and titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.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

 مرئيه

-0.54

IntoConstraints

-0.52

httphttps

-0.47

Mau

-0.45

Mast

-0.44

 Mast

-0.44

 mxArray

-0.43

ValueStyle

-0.42

 للمعارف

-0.41

 пости

-0.41

POSITIVE LOGITS

 Delta

0.48

 DELTA

0.42

Delta

0.42

 Administrativna

0.40

 nisso

0.40

Rüyada

0.39

 creeks

0.39

 rövid

0.39

 delta

0.38

Derbyniad

0.38

Activations Density 0.127%

proper nouns, specifically names and titles

No Comments

No Known Activations