INDEX

Explanations

references to specific aims and goals within scientific or structured proposals

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

labels and list items

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

 للمعارف

-0.82

 Италијани

-0.78

 Taktlose

-0.72

 мәкал

-0.69

tagHelperRunner

-0.69

 gynhyrchwyd

-0.67

 OMITBAD

-0.66

OGND

-0.66

 pinulongan

-0.66

Autoritní

-0.65

POSITIVE LOGITS

Câu

0.42

Bài

0.42

Bước

0.41

Chapter

0.40

Figure

0.40

package

0.40

↵↵

0.40

The

0.40

 Fort

0.38

0.37

Activations Density 0.064%

references to specific aims and goals within scientific or structured proposals

labels and list items

No Comments

No Known Activations