INDEX

Explanations

references to statements and formal communications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

IUrlHelper

-0.43

 neceff

-0.33

 proje

-0.32

zás

-0.32

nezeu

-0.32

 fréqu

-0.32

 dépens

-0.32

 Jefus

-0.32

 rencontr

-0.31

 cierta

-0.31

POSITIVE LOGITS

 letter

0.73

 defaultstate

0.70

 written

0.70

 lettera

0.64

PMailer

0.61

parsedMessage

0.60

 epistle

0.59

written

0.58

 penned

0.56

+#+#

0.56

Activations Density 0.630%

references to statements and formal communications

No Comments

No Known Activations