INDEX

Explanations

references to player interactions and game mechanics

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

user input with game terms

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

DoubleQuotes

-0.39

锈钢

-0.37

stdc

-0.35

__':

-0.35

ře

-0.35

xbd

-0.34

 arşivlendi

-0.33

houding

-0.33

__':

-0.33

 JADX

-0.33

POSITIVE LOGITS

 CreateTagHelper

0.63

fromnode

0.61

 getItemId

0.56

 jsPsych

0.51

arium

0.48

 NSCoder

0.47

LookAnd

0.47

 betweenstory

0.46

Carriera

0.46

 gameplay

0.45

Activations Density 0.053%

references to player interactions and game mechanics

user input with game terms

No Comments

No Known Activations