INDEX

Explanations

database-related terms and configurations

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

passwords after numbers/old

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

AndEndTag

-0.90

MLLoader

-0.80



-0.79

 للاسماء

-0.75

 GenerationType

-0.75

<unused28>

-0.73

<pad>

-0.73

<unused16>

-0.73

[@BOS@]

-0.73

<unused8>

-0.73

POSITIVE LOGITS

PasswordEncoder

0.24

0.23

 Monats

0.23

 faciles

0.23

 boisson

0.22

 Rückseite

0.22

 diseños

0.21

 secas

0.20

 tausend

0.20

lib

0.20

Activations Density 0.005%

database-related terms and configurations

passwords after numbers/old

No Comments

No Known Activations