INDEX

Explanations

technical language related to game mechanics and spells

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

some may

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

IntoConstraints

-0.43

 pinulongan

-0.40

信

-0.39

_),

-0.36

 autorytatywna

-0.35

IBOutlet

-0.34

 réessayer

-0.34

 spørgsmål

-0.34

 linkovi

-0.33

 ostavi

-0.33

POSITIVE LOGITS

RegressionTest

0.72

TextStyle

0.50

blurRadius

0.50

Segoe

0.48

piram

0.48

 太郎

0.47

 IconData

0.46

 ItemStack

0.45

carre

0.45

 aust

0.44

Activations Density 0.051%

technical language related to game mechanics and spells

some may

No Comments

No Known Activations