INDEX

Explanations

terms related to structured data and formal documents

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

Bash built, CT scan, at coordinates

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

 Wiktionnaire

-0.61

AndEndTag

-0.60

jenost

-0.59



-0.52

 noDo

-0.49

 Inscrivez

-0.49

 utafitiHapana

-0.49

 Privacidade

-0.48

 bambú

-0.48

 ujednoznacz

-0.47

POSITIVE LOGITS

ArrowToggle

0.38

蚪

0.34

Hy

0.33

isContained

0.32

Zip

0.32

Tembelea

0.32

 metas

0.31

بوابة

0.31

hy

0.30

&&\

0.30

Activations Density 0.121%

terms related to structured data and formal documents

Bash built, CT scan, at coordinates

No Comments

No Known Activations