INDEX

Explanations

proper nouns, particularly names and titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_11/width_16k/average_l0_118

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

antaranya

-0.52

 Administrativna

-0.47

 disambiguazione

-0.47

 terdekat

-0.46

 siihen

-0.46

 nemlig

-0.45

 moeite

-0.45

 prohibido

-0.44

 osallist

-0.44

 stør

-0.43

POSITIVE LOGITS

 dAtA

0.54

 ***!

0.53

ScopeManager

0.48

="@+

0.46

WithIOException

0.45

 gubernatorial

0.45

 institution

0.42

 backups

0.42

Prog

0.42

Theo

0.41

Activations Density 0.928%

proper nouns, particularly names and titles

No Comments

No Known Activations