INDEX

Explanations

terms related to code structure and graphical models

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

technical tokens from code, math/LaTeX, and scientific writing—especially abbreviations, symbols, and hyphenated or underscored forms.

oai_token-act-pair · gpt-5 Triggered by @arielgoldsteinlab

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

 Wikiseite

-0.28

 déclar

-0.27

 this

-0.27

 ferons

-0.26

 présentes

-0.26

camiseta

-0.26

 particuliers

-0.25

 habr

-0.25

 marqués

-0.25

 choisissez

-0.25

POSITIVE LOGITS

OGND

0.96

HostException

0.96

 مشين

0.89

ScopeManager

0.88

存于互联网档案馆

0.86

 esternos

0.79

Geplaatst

0.76

FormTagHelper

0.75

enfone

0.74

 ivelany

0.74

Activations Density 2.695%

terms related to code structure and graphical models

technical tokens from code, math/LaTeX, and scientific writing—especially abbreviations, symbols, and hyphenated or underscored forms.

No Comments

No Known Activations