INDEX

Explanations

numeric references, particularly years and possibly other quantitative data

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

numbers and years

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

httphttps

-0.51

 defaultstate

-0.49

 nahilalakip

-0.48

ChildScrollView

-0.47

存于互联网档案馆

-0.45

+#+

-0.45

yntaxException

-0.45

ſelf

-0.44

Jeografia

-0.44

ſelves

-0.43

POSITIVE LOGITS

 Foundation

0.48

הערות

0.46

 ۱۳۹

0.45

 endblock

0.42

昭和

0.41

AppMethodBeat

0.41

 Walden

0.40

parseColor

0.40

Foundation

0.40

平成

0.40

Activations Density 0.010%

numeric references, particularly years and possibly other quantitative data

numbers and years

No Comments

No Known Activations