INDEX

Explanations

references to national-related terms and concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 3-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_3/width_16k/average_l0_59

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.3.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

 Assad

-0.66

ECB

-0.63

 Gerv

-0.62

BITDA

-0.62

 YOGA

-0.60

 archaeologist

-0.60

urbed

-0.59

 Lyft

-0.59

 Catto

-0.59

handelt

-0.58

POSITIVE LOGITS

 national

3.21

national

2.83

 National

2.60

National

2.48

NATIONAL

2.42

 NATIONAL

2.35

 nationale

2.25

 nacional

2.24

 nationally

2.07

 nationalen

1.98

Activations Density 0.052%

references to national-related terms and concepts

No Comments

No Known Activations