INDEX

Explanations

elements that denote structure or formatting in a programming or markup context

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

math and programming code

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

 اف

-0.32

"__

-0.31

再

-0.30

政

-0.29

ante

-0.28

 للا

-0.28

الت

-0.27

 allá

-0.27

 fuerzas

-0.26

asi

-0.26

POSITIVE LOGITS

 ſind

0.75

 للمعارف

0.72

httphttps

0.71

 queſta

0.67

mybatisplus

0.66

ſchaft

0.64

PMailer

0.64

 ddelwed

0.62

TagMode

0.62

 kasarigan

0.62

Activations Density 0.001%

elements that denote structure or formatting in a programming or markup context

math and programming code

No Comments

No Known Activations