INDEX

Explanations

references to private/internal functions or variables in code

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

internal memos

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

featureID

-0.69

ChildScrollView

-0.63

SharedCtor

-0.63

 Paglinawan

-0.57

MessageOf

-0.54

IntoConstraints

-0.50

EDEFAULT

-0.48

 ModelRenderer

-0.48

 Photocase

-0.47

Vtbl

-0.45

POSITIVE LOGITS

Derbyniad

0.42

ForTesting

0.42

内部

0.40

Примітки

0.39

 Orts

0.36

ANDUM

0.36

internal

0.35

Билгалдахарш

0.35

wußt

0.35

ServerError

0.34

Activations Density 0.016%

references to private/internal functions or variables in code

internal memos

No Comments

No Known Activations