INDEX

Explanations

numerical values or identifiers related to data and measurements

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

German words

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

 sebou

-0.35

PreferredItem

-0.34

elemField

-0.34

 picioare

-0.34

 înal

-0.33

 kullanılan

-0.33

 anciennes

-0.33

 niej

-0.33

 nisso

-0.33

 scorso

-0.32

POSITIVE LOGITS

 zuſammen

0.71

 queſta

0.68

 zwiſchen

0.68

 ſind

0.66

 ſei

0.65

 ſehr

0.65

 geſch

0.65

存于互联网档案馆

0.65

ſchaft

0.63

OCCURRED

0.62

Activations Density 0.034%

numerical values or identifiers related to data and measurements

German words

No Comments

No Known Activations