INDEX

Explanations

references to physical space or areas within contexts

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

space

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

<bos>

-0.49

 tunggal

-0.42

Doktor

-0.40

Try

-0.39

-0.38

the

-0.36

 Doktor

-0.35

 conclusão

-0.35

 Third

-0.34

 zdan

-0.34

POSITIVE LOGITS

Space

1.28

 SPACE

1.27

 space

1.24

SPACE

1.24

 Space

1.23

Spaces

1.20

 Spaces

1.19

space

1.14

itespace

1.12

 spaces

1.06

Activations Density 0.021%

references to physical space or areas within contexts

space

No Comments

No Known Activations