INDEX

Explanations

alphanumeric sequences that appear to be identifiers or codes

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

 Chwiliwch

-0.60

angliski

-0.50

EndGlobalSection

-0.49

 himſelf

-0.47

 становника

-0.45

 appunto

-0.44

دانشنامهٔ

-0.43

 GenerationType

-0.43

libft

-0.42

Glej

-0.42

POSITIVE LOGITS

 estekak

0.55

aclk

0.41

 otomatig

0.40

 uVar

0.40

 MenuView

0.38

PYX

0.38

 defaultstate

0.36

期刊论文

0.36

 SqlCommand

0.36

Diwedd

0.35

Activations Density 0.255%

alphanumeric sequences that appear to be identifiers or codes

No Comments

No Known Activations