INDEX

Explanations

references to specific social groups and roles

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

user groups

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

 виправивши

-0.68

 RIPRODUZIONE

-0.46

 alcanzado

-0.42

Commentaires

-0.42

__':

-0.42

 esegu

-0.42

 Grüßen

-0.42

 arreglos

-0.40

 kasarigan

-0.40

nessed

-0.40

POSITIVE LOGITS

 számára

0.48

__*/

0.46

 investors

0.46

yyl

0.45

 beginners

0.44

 עוב

0.44

mtable

0.43

YL

0.43

 skiers

0.43

 lawyers

0.42

Activations Density 0.043%

references to specific social groups and roles

user groups

No Comments

No Known Activations