INDEX

Explanations

words and phrases associated with specific categories or classifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

 Keuangan

-1.17

 Masyarakat

-1.09

 Budaya

-1.08

 Lingkungan

-1.07

 myſelf

-1.04

 Penga

-1.02

 Kebijakan

-1.02

 Komunikasi

-1.01

 Manusia

-1.01

 houſe

-1.01

POSITIVE LOGITS

na

0.63

pod

0.63

od

0.63

pro

0.61

roz

0.60

vy

0.60

bez

0.59

„

0.57

tak

0.57

Activations Density 0.054%

words and phrases associated with specific categories or classifications

No Comments

No Known Activations