INDEX

Explanations

code structures, particularly class declarations and components in programming

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

 Мексичка

-0.94

 Taktlose

-0.85

تقاوى

-0.79

 Administrativna

-0.73

 autorytatywna

-0.73

 gyhoeddwyd

-0.70

contentLoaded

-0.69

thâu

-0.69

 Infórmanos

-0.68

#+#

-0.68

POSITIVE LOGITS

 Master

0.33

 model

0.32

볍

0.31

特点

0.31

 Manager

0.31

元年

0.30

 мето

0.29

Act

0.29

 Dimension

0.29

Heights

0.29

Activations Density 0.627%

code structures, particularly class declarations and components in programming

No Comments

No Known Activations