INDEX

Explanations

references to historical events and figures

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

years and dates

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

دانشنامهٔ

-0.59

 CanadaChoose

-0.59

OPHER

-0.49

 CAND

-0.49

 AssemblyTitle

-0.48

 nuclear

-0.48

 ModelExpression

-0.47

 Normdatei

-0.46

omock

-0.46

 Obama

-0.46

POSITIVE LOGITS

Victorian

0.47

 ferrocarril

0.44

 enfans

0.44

kheim

0.42

 žele

0.42

 itſelf

0.41

 industriels

0.40

ITURE

0.40

 poffible

0.40

 fotográfico

0.40

Activations Density 0.134%

references to historical events and figures

years and dates

No Comments

No Known Activations