INDEX

Explanations

references to specific characteristics or details in discussions or narratives

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

 Biôgrafia

-0.69

ésultats

-0.61

síða

-0.60

WriteAttribute

-0.59

twimg

-0.59

 Waſſer

-0.56

 Paglinawan

-0.55

writeFieldEnd

-0.55

iffance

-0.54

يكب

-0.54

POSITIVE LOGITS

‌است

0.29

اریخ

0.29

CppMethod

0.28

 vorge

0.28

haltung

0.28

Locator

0.27

到的

0.27

lorin

0.26

Dtor

0.26

 incomp

0.26

Activations Density 0.204%

references to specific characteristics or details in discussions or narratives

No Comments

No Known Activations