INDEX

Explanations

phrases indicating the beginning or introduction to a larger idea or concept

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

only one of

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

 Breaks

-0.47

 Break

-0.47

出版年

-0.46

Rüyada

-0.43

üyada

-0.40

CreateModel

-0.40

 BREAK

-0.38

 breaks

-0.38

break

-0.38

RTLU

-0.38

POSITIVE LOGITS

罷

0.54

EndContext

0.45

罢了

0.43

 onderdeel

0.43

而已

0.41

chit

0.40

 augmentation

0.40

࿐

0.40

 HasFactory

0.40

 oprot

0.40

Activations Density 0.029%

phrases indicating the beginning or introduction to a larger idea or concept

only one of

No Comments

No Known Activations