INDEX

Explanations

technical terms and precise measurements related to data analysis and processing

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

nearest decimal or resolution

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

⟬

-0.44

OGND

-0.38

MLLoader

-0.36

TextAppearance

-0.35

锈钢

-0.34

CUM

-0.33

 packs

-0.32

员

-0.32

 OMITBAD

-0.32

 Cæsar

-0.32

POSITIVE LOGITS

 pouce

0.50

Semoga

0.49

 mịn

0.48

 gouttes

0.47

Tembelea

0.47

Добрый

0.44

 pohy

0.44

Resolution

0.43

 evitando

0.43

 fluide

0.43

Activations Density 0.271%

technical terms and precise measurements related to data analysis and processing

nearest decimal or resolution

No Comments

No Known Activations