INDEX

Explanations

specific entities, such as organizations or locations, indicated by proper nouns

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

devices 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

Launch

-0.33

 Launch

-0.33

launch

-0.33

saraba

-0.32

skin

-0.32

fancy

-0.31

 aceptas

-0.31

 launch

-0.30

 LAUNCH

-0.30

Soil

-0.30

POSITIVE LOGITS

tonsoft

0.52

GEBURTSDATUM

0.51

UnusedPrivate

0.51

readObject

0.49

 חיצוניים

0.49

parsedMessage

0.48

SharedDtor

0.47

TypedDataSet

0.47

Зноскі

0.46

ValueStyle

0.43

Activations Density 0.237%

specific entities, such as organizations or locations, indicated by proper nouns

devices and questions

No Comments

No Known Activations