INDEX

Explanations

names of people and their affiliations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

middels

-0.50

tershire

-0.45

彿

-0.42

 ANGELES

-0.42

intenant

-0.41

 noDo

-0.41

 brooks

-0.41

ỡng

-0.41

 gynhyrchwyd

-0.39

 airs

-0.38

POSITIVE LOGITS

 surla

0.69

 died

0.53

 murieron

0.48

 posthum

0.47

 founding

0.43

 murió

0.43

这个人

0.42

 famous

0.41

EndContext

0.40

 genius

0.40

Activations Density 0.541%

names of people and their affiliations

No Comments

No Known Activations