INDEX

Explanations

phrases emphasizing exceptional qualities or rankings, particularly in relation to people, places, and categories

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

eschön

-0.48

তথ্যসূত্র

-0.46

 Decke

-0.45

astify

-0.44

 Ganzen

-0.44

xaml

-0.43

 Tatsache

-0.43

 Füße

-0.43

 daß

-0.42

 gesamte

-0.42

POSITIVE LOGITS

GEBURTSDATUM

0.66

 among

0.65

 Among

0.61

among

0.60

Among

0.60

 highly

0.58

 AMONG

0.55

 Amongst

0.54

 Савезне

0.54

 toughest

0.53

Activations Density 0.146%

phrases emphasizing exceptional qualities or rankings, particularly in relation to people, places, and categories

No Comments

No Known Activations