INDEX

Explanations

proper nouns related to institutions and medical terms

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_11/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 crin

-0.41

setLong

-0.37

bry

-0.36

 TOGETHER

-0.35

 impost

-0.35

rescue

-0.34

 regn

-0.34

Lon

-0.34

Resumen

-0.33

ItemBackground

-0.33

POSITIVE LOGITS

 Koreans

0.64

CppMethod

0.60

 Korea

0.57

Korea

0.56

 estekak

0.55

LEncoder

0.55

Korean

0.55

Хьажоргаш

0.54

 Korean

0.54

tanleria

0.53

Activations Density 0.827%

proper nouns related to institutions and medical terms

No Comments

No Known Activations