INDEX

Explanations

references to specific websites and numerical identifiers

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

own, native, or belonging

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

rungsseite

-0.84

 utafitiHapana

-0.80

 Signalez

-0.73

ChildScrollView

-0.72

 CanadaChoose

-0.68

+#+#

-0.68

 queſta

-0.67

曖昧さ回避

-0.65

posedge

-0.65

Datuak

-0.65

POSITIVE LOGITS

 propios

0.45

own

0.45

 native

0.40

 propia

0.39

 vlast

0.39

 própria

0.39

 próprios

0.38

 propio

0.37

 eigene

0.37

 propias

0.35

Activations Density 0.155%

references to specific websites and numerical identifiers

own, native, or belonging

No Comments

No Known Activations