INDEX

Explanations

phrases related to narrative techniques and authorial intent

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

fact gain, may take

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

打

-0.32

 Waray

-0.31

endregion

-0.30

tagHelperRunner

-0.30

 olet

-0.30

aarrggbb

-0.30

sak

-0.30

 Vereinigte

-0.29

 AssemblyCompany

-0.29

 Sekret

-0.29

POSITIVE LOGITS

 فريبيس

0.62

 ſind

0.61

 beſte

0.59

 queſta

0.59

$_"

0.59

 Chwiliwch

0.59

 sikkert

0.57

niſſe

0.57

ſelben

0.56

ſchaft

0.56

Activations Density 0.117%

phrases related to narrative techniques and authorial intent

fact gain, may take

No Comments

No Known Activations