INDEX

Explanations

specific gene names and molecular terms associated with biological processes

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

prefixes of multi-word names

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

 CanadaChoose

-0.73

OGND

-0.69

adaptiveStyles

-0.69

Autoritní

-0.68

 autorytatywna

-0.68

Tembelea

-0.65

цездатний

-0.65

يكب

-0.64

BagConstraints

-0.63

 tartalomajánló

-0.60

POSITIVE LOGITS

is

0.60

has

0.49

 also

0.49

The

0.49

 doesn

0.46

can

0.43

 stessi

0.43

may

0.42

 buono

0.41

Activations Density 0.064%

specific gene names and molecular terms associated with biological processes

prefixes of multi-word names

No Comments

No Known Activations