INDEX

Explanations

references to quantity or numerical expressions

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

quantifiers and counting words

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 9-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_9/width_131k/average_l0_121

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.9.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

 mijne

-0.57

setopt

-0.43

arithmic

-0.42

šech

-0.41

 Inscrivez

-0.41

 identitas

-0.40

setDo

-0.40

 quæ

-0.39

 oorlog

-0.39

σθαι

-0.39

POSITIVE LOGITS

兩個

0.87

一个

0.83

两个

0.83

三個

0.80

 一个

0.79

幾個

0.77

三个

0.75

一個

0.74

第二个

0.73

几个

0.69

Activations Density 0.003%

references to quantity or numerical expressions

quantifiers and counting words

No Comments

No Known Activations