INDEX

Explanations

references to variables and data structures

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

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 υπό

-0.36

didSet

-0.35

точно

-0.35

 <<<<<<<<<<<<<<

-0.34

 ferons

-0.34

.');

-0.33

"")

-0.33

sto

-0.32

 HOLL

-0.32

vig

-0.32

POSITIVE LOGITS

expandindo

0.62

CloseOperation

0.57

styleType

0.57

 MenuView

0.55

rungsseite

0.54

 تضيفلها

0.53

ագրություններ

0.53

 noDo

0.53

 Himo

0.53

 виправивши

0.53

Activations Density 0.059%

references to variables and data structures

No Comments

No Known Activations