INDEX

Explanations

questioning phrases directed at confirming or evaluating statements and values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 12-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_12/width_16k/average_l0_130

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.12.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

ſelf

-0.44

 ſta

-0.44

 faſt

-0.44

 myſelf

-0.44

ſelves

-0.44

 internetowa

-0.42

ьаж

-0.42

 queſta

-0.40

/*

-0.39

éez

-0.39

POSITIVE LOGITS

 ujednoznacz

0.51

IBOutlet

0.46

 surla

0.44

rungsseite

0.44

mel

0.42

airbnb

0.42

lau

0.41

detectChanges

0.40

SequentialGroup

0.40

 onOptions

0.40

Activations Density 0.597%

questioning phrases directed at confirming or evaluating statements and values

No Comments

No Known Activations