INDEX

Explanations

numeric patterns or sequences in the text

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

start of user turn

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

 Biôgrafia

-0.54

fel

-0.44

ladel

-0.40

逐

-0.40

NSIndexPath

-0.40

enderror

-0.39

erd

-0.39

malink

-0.39

asson

-0.39

TZ

-0.39

POSITIVE LOGITS

tagHelperRunner

0.57

afficheront

0.48

 فريبيس

0.47

Tikang

0.46

NameInMap

0.45

 restTemplate

0.45

 neón

0.45

 zijne

0.44

Personensuche

0.44

 autorytatywna

0.43

Activations Density 0.002%

numeric patterns or sequences in the text

start of user turn

No Comments

No Known Activations