INDEX

Explanations

elements that denote mathematical or logical relationships and structures

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

Mathematical/logical expressions and equations

np_acts-logits-general · gemini-2.0-flash

state unchanged

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

Personensuche

-0.73

expandindo

-0.62

ReusableCell

-0.60

TypedDataSet

-0.54

 PeEnEo

-0.54

urlencoded

-0.53

UnusedPrivate

-0.52

 StatelessWidget

-0.51

 архивлан

-0.50

 OFDb

-0.48

POSITIVE LOGITS

 tetap

0.62

 vẫn

0.58

 remain

0.56

 unchanged

0.56

依旧

0.56

変わらず

0.53

 remains

0.52

 nadal

0.49

 continue

0.49

 unaltered

0.48

Activations Density 0.572%

elements that denote mathematical or logical relationships and structures

Mathematical/logical expressions and equations

state unchanged

No Comments

No Known Activations