INDEX

Explanations

strong descriptors and phrases emphasizing quality or positivity

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

positive assessments

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

 Équipe

-0.38

dø

-0.37

 Occidente

-0.35

 émer

-0.35

 pleaſure

-0.34

 Conſ

-0.33

 flamme

-0.32

ßerdem

-0.32

 actionPerformed

-0.32

 Polsek

-0.32

POSITIVE LOGITS

 للاسماء

0.73

 juicy

0.66

OGND

0.65

 good

0.63

 hearty

0.62

BufferException

0.61

 BoxFit

0.59

 ***!

0.58

bootstrapcdn

0.58

Personensuche

0.57

Activations Density 0.013%

strong descriptors and phrases emphasizing quality or positivity

positive assessments

No Comments

No Known Activations