INDEX

Explanations

punctuation marks, particularly exclamation points and question marks, indicating strong emotional or interrogative expressions

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

exclamations and questions

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

 avoient

-0.62

 auroit

-0.52

 étoit

-0.51

 pouvoit

-0.51

 feroit

-0.51

 nôtre

-0.50

 étranger

-0.50

 étrangère

-0.50

 étoient

-0.46

 chrétien

-0.45

POSITIVE LOGITS

 ་་

0.74

Билгалдахарш

0.73

 Zeppelin

0.70

يكب

0.69

 Clarion

0.69

figuration

0.69

 Starship

0.68

 Taktlose

0.68

Santis

0.66

GLS

0.66

Activations Density 0.009%

punctuation marks, particularly exclamation points and question marks, indicating strong emotional or interrogative expressions

exclamations and questions

No Comments

No Known Activations