INDEX

Explanations

references to the concept of "self" or self-referential constructs

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

itself

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

aarrggbb

-0.57

']))

-0.54

MessageTagHelper

-0.52

 الاطلاع

-0.51

arithmic

-0.50

UnitTesting

-0.50

 Kangaroo

-0.49

ReusableCell

-0.49

</>

-0.49

 Paralympic

-0.49

POSITIVE LOGITS

 itself

0.91

itself

0.88

 Itself

0.84

 kendisi

0.54

 Himself

0.53

 sebenarnya

0.50

himself

0.50

 himself

0.48

 själv

0.48

ebenarnya

0.48

Activations Density 0.016%

references to the concept of "self" or self-referential constructs

itself

No Comments

No Known Activations