INDEX

Explanations

references to measurements and quantitative data

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

area and dimensions

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

internalType

-0.33

 skeleton

-0.30

gefahr

-0.30

 strap

-0.29

 Stä

-0.28

 gond

-0.28

 flan

-0.28

 rend

-0.28

peicher

-0.28

kö

-0.28

POSITIVE LOGITS

area

0.63

AREA

0.61

Area

0.60

ChildScrollView

0.60

 Area

0.60

areas

0.59

 Chwiliwch

0.59

面積

0.58

 áreas

0.58

 área

0.58

Activations Density 0.112%

references to measurements and quantitative data

area and dimensions

No Comments

No Known Activations