INDEX

Explanations

phrases related to devices, organizations, and situations particularly noted for their significance or requirements

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

 Erschein

-0.38

 تضيفلها

-0.36

harusnya

-0.33

 dificuldades

-0.33

amsmath

-0.32

 limitada

-0.32

 المعيارى

-0.32

GraphicsUnit

-0.31

 Konflikt

-0.30

 signé

-0.30

POSITIVE LOGITS

 autorytatywna

0.80

 transfieras

0.63

ſelf

0.61

UserScript

0.60

 Verſ

0.60

 onOptions

0.57

 itſelf

0.56

 ſch

0.56

expandindo

0.56

 ſche

0.55

Activations Density 0.192%

phrases related to devices, organizations, and situations particularly noted for their significance or requirements

No Comments

No Known Activations