INDEX

Explanations

names of people and their affiliations

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

names or initials followed by titles

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

 kasarigan

-0.99

astify

-0.85

 AssemblyCulture

-0.85

 المعيارى

-0.80

 EconPapers

-0.80

ロウィン

-0.74

 autorytatywna

-0.74

iſche

-0.71

 intptr

-0.71

OGND

-0.71

POSITIVE LOGITS

上线

0.30

上

0.29

Integer

0.29

usiai

0.28

This

0.28

tiennent

0.27

 أس

0.26

<strong>

0.26

städ

0.26

Activations Density 0.104%

names of people and their affiliations

names or initials followed by titles

No Comments

No Known Activations