INDEX

Explanations

phrases that discuss or evaluate quality and quantity of various subjects

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

quality of

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 utafitiHapana

-0.38

expand

-0.37

 Millar

-0.36

 تعدى

-0.36

thren

-0.36

 violación

-0.35

RenderAtEndOf

-0.35

 Umum

-0.35

 Lieben

-0.35

 stub

-0.35

POSITIVE LOGITS

 quality

1.27

quality

1.21

Quality

1.16

 Quality

1.10

 QUALITY

1.02

 qualidade

0.99

 Qualität

0.98

Qualität

0.95

 kwaliteit

0.94

QUALITY

0.93

Activations Density 0.022%

phrases that discuss or evaluate quality and quantity of various subjects

quality of

No Comments

No Known Activations