INDEX

Explanations

words indicating quantity or frequency

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

describing properties

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

 esprits

-0.47

englanniksi

-0.41

 Brace

-0.39

 ajan

-0.38

PerformLayout

-0.35

AssignableFrom

-0.34

fontStyle

-0.34

 wijze

-0.33

 acaso

-0.32

Hãy

-0.32

POSITIVE LOGITS

 gehad

0.60

 Normdatei

0.59

 surla

0.56

 Tiene

0.56

TagMode

0.56

 propOrder

0.54

principalColumn

0.54

hasMoreElements

0.52

 հղումներ

0.50

 habet

0.50

Activations Density 0.081%

words indicating quantity or frequency

describing properties

No Comments

No Known Activations