INDEX

Explanations

discussions about personal or social interactions

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

preceding personal pronouns

np_acts-logits-general · gemini-2.0-flash

people or individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

IonicModule

-0.65

 Roskov

-0.63

AddHtmlAttribute

-0.61

HasAnnotation

-0.60

ijnlijk

-0.59

IMPORTED

-0.58

IntoConstraints

-0.58

󠁢

-0.57

 ProtoMessage

-0.57

 Chwiliwch

-0.56

POSITIVE LOGITS

 people

1.36

 everyone

1.14

people

1.11

 People

1.11

People

1.09

 everybody

1.05

 nobody

1.00

人们

0.98

 PEOPLE

0.98

everyone

0.97

Activations Density 0.455%

discussions about personal or social interactions

preceding personal pronouns

people or individuals

No Comments

No Known Activations