INDEX

Explanations

references to specific individuals, particularly focusing on 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 @ 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



-0.54

 banques

-0.47

 Komunikasi

-0.45

 navidad

-0.45

 demikian

-0.44

 exceptionnel

-0.43

 Latinoamérica

-0.43

 tiegħ

-0.43

 Dodson

-0.43

twimg

-0.43

POSITIVE LOGITS

SourceChecksum

0.54

 يتيمه

0.49

➟

0.47

yargs

0.47

Personendaten

0.43

 الحره

0.42

[''

0.42

⁡

0.42

 kasarigan

0.41

avits

0.41

Activations Density 0.188%

references to specific individuals, particularly focusing on names and titles

No Comments

No Known Activations