INDEX

Explanations

references to scientific articles or studies

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

doi codes

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

KommentareTeilen

-0.62

 surla

-0.60

 AssemblyTitle

-0.56

 iPod

-0.55

iPod

-0.52

 kasarigan

-0.52

balleur

-0.50

 MySpace

-0.50

 createSprite

-0.48

osene

-0.47

POSITIVE LOGITS

 COVID

0.58

COVID

0.52

令和

0.52

StandardCharsets

0.52

 Covid

0.50

 covid

0.48

コロナ

0.47

covid

0.47

0.46

Covid

0.45

Activations Density 0.187%

references to scientific articles or studies

doi codes

No Comments

No Known Activations