INDEX

Explanations

references to historical or religious figures and places

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

ſelves

-0.49

multer

-0.46

 poffe

-0.46

 compareTo

-0.45

 سكانية

-0.45

Og

-0.43

 vpon

-0.43

ientras

-0.42

 يتيمه

-0.42

 fevere

-0.42

POSITIVE LOGITS

 richest

0.48

 saddest

0.44

 galería

0.43

ंदीखरीदारी

0.42

XmlAccessorType

0.42

 toughest

0.41

 hardest

0.41

 slowest

0.40

 darkest

0.40

IVEREF

0.40

Activations Density 0.062%

references to historical or religious figures and places

No Comments

No Known Activations