INDEX

Explanations

specific programming-related terms, particularly those associated with threads, images, and object-oriented programming concepts

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

start_of_turn user

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

󠁢

-0.69

 esimer

-0.49

Билгалдахарш

-0.46

 BoxFit

-0.45

NameInMap

-0.44

legte

-0.41

posedge

-0.41

 tärke

-0.40

 devrez

-0.40

ised

-0.40

POSITIVE LOGITS

FormTagHelper

0.51

endphp

0.50

 council

0.46

 commission

0.44

 authorities

0.44

contentPane

0.43

 company

0.42

 isNameExpr

0.41

 supplier

0.41

 Commission

0.39

Activations Density 0.002%

specific programming-related terms, particularly those associated with threads, images, and object-oriented programming concepts

start_of_turn user

No Comments

No Known Activations