INDEX

Explanations

names and possibly some specific types of proper nouns

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

Prefixes of names or words

np_acts-logits-general · gemini-2.0-flash

academic and legal citations

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

abello

-0.71

setVerticalGroup

-0.69

orini

-0.67

glLoadIdentity

-0.65

PerformLayout

-0.63

ValueStyle

-0.61

oldi

-0.58

AnimationsModule

-0.58

haupt

-0.57

Véxase

-0.56

POSITIVE LOGITS

MBE

0.64

DMD

0.63

 rospy

0.62

ⓧ

0.61

PHD

0.60

 Klan

0.59

 StatefulWidget

0.58

 egret

0.58

 RIPRODUZIONE

0.58

HasAnnotation

0.57

Activations Density 0.286%

names and possibly some specific types of proper nouns

Prefixes of names or words

academic and legal citations

No Comments

No Known Activations