INDEX

Explanations

numerical values and configurations related to programming or game development

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

numbers in code or math

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

BeginContext

-0.79

 ſind

-0.75

 EconPapers

-0.73

UrlResolution

-0.72

المناصب

-0.71

 ſche

-0.67

 وتسجيلات

-0.66

الإنجليزية

-0.65

ðsíða

-0.64

ImageContext

-0.63

POSITIVE LOGITS

 darte

0.35

outSlope

0.33

 asegurarse

0.32

 corriendo

0.32

 esperamos

0.31

 sabido

0.31

OrderDetail

0.31

 inmediatamente

0.31

po

0.30

 mesmo

0.30

Activations Density 0.188%

numerical values and configurations related to programming or game development

numbers in code or math

No Comments

No Known Activations