INDEX

Explanations

conditional statements or phrases indicating hypothetical scenarios

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

as if introducing comparison

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

 órgano

-0.58

AndEndTag

-0.54

 courants

-0.54

 gesloten

-0.53

 giudizio

-0.50

 termica

-0.49

erals

-0.49

 pescoço

-0.49

 wortel

-0.48

তথ্যসূত্র

-0.47

POSITIVE LOGITS

GEBURTSDATUM

0.43

 jakby

0.41

ViewInit

0.40

껏

0.39

URIComponent

0.39

strptime

0.38

терак

0.38

sendKeys

0.37

 elseif

0.37

 aligning

0.36

Activations Density 0.003%

conditional statements or phrases indicating hypothetical scenarios

as if introducing comparison

No Comments

No Known Activations