INDEX

Explanations

questions starting with "how" and their corresponding answers

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

followed by conjugations of "be" or "do"

np_acts-logits-general · gemini-2.0-flash

how something works or differs

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 18-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_18/width_16k/average_l0_74

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.18.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

certainly

-0.59

LookAnd

-0.57

WebElementEntity

-0.56

 certainly

-0.54

 defaultstate

-0.52

definitely

-0.51

defn

-0.50

 solides

-0.48

chtend

-0.46

SequentialGroup

-0.46

POSITIVE LOGITS

fjspx

0.77

 differs

0.76

 snippetHide

0.75

 compares

0.72

 differed

0.69

 Became

0.68

 differ

0.68

 became

0.68

 ویکی‌پدیا

0.65

SharedDtor

0.64

Activations Density 0.176%

questions starting with "how" and their corresponding answers

followed by conjugations of "be" or "do"

how something works or differs

No Comments

No Known Activations

questions starting with "how" and their corresponding answers

followed by conjugations of "be" or "do"

how something works or differs

No Comments

No Known Activations