INDEX

Explanations

programming terms and elements related to matrix operations and structure

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

ImplTest class names

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

AddTagHelper

-0.86

 MainAxisSize

-0.82

 للاسماء

-0.77

Tikang

-0.76

RectangleBorder

-0.73

AsUp

-0.72

FunctionFlags

-0.71

parsedMessage

-0.70

RotationOrder

-0.68

featureID

-0.67

POSITIVE LOGITS

world

0.33

 Gubernur

0.31

0.30

 Ensino

0.30

 Schae

0.29

 game

0.28

世界

0.27

 scenery

0.27

World

0.27

Activations Density 0.035%

programming terms and elements related to matrix operations and structure

ImplTest class names

No Comments

No Known Activations