INDEX

Explanations

sequences of code or programming constructs

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

final concentration

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

DrawerToggle

-0.40

 handicap

-0.37

 actionMode

-0.36

PUL

-0.35

 wireType

-0.34

爆

-0.34

ae

-0.34

COP

-0.33

pul

-0.33

Ant

-0.33

POSITIVE LOGITS

 Taktlose

0.71

 geweest

0.62

 genieten

0.59

 geschiedenis

0.59

üyada

0.57

 zijne

0.56

 bedoeld

0.56

 autorytatywna

0.56

 mêmes

0.55

المناصب

0.55

Activations Density 0.109%

sequences of code or programming constructs

final concentration

No Comments

No Known Activations