INDEX

Explanations

references to teddy bears and stuffed animals

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

teddy bear

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

 AttributeSet

-0.56

RUnlock

-0.54

:✨

-0.52

seido

-0.51

IContainer

-0.51

yonel

-0.48

loem

-0.48

atimes

-0.47

 TestBed

-0.46

Traducción

-0.46

POSITIVE LOGITS

 cuddly

0.43

 plush

0.42

igurumi

0.42

Trayectoria

0.40

这只

0.38

 cuddle

0.37

 amigurumi

0.36

 oversized

0.35

 Plush

0.34

 peluche

0.33

Activations Density 0.006%

references to teddy bears and stuffed animals

teddy bear

No Comments

No Known Activations