INDEX

Explanations

phrases indicating addition or elaboration on previous points

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

another or also plus description

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

sizeCache

-0.50

CROSSTALK

-0.43

اریخ

-0.42

HasAnnotation

-0.42

 Roskov

-0.42

PropertyChanging

-0.41

 själ

-0.41

aarrggbb

-0.40

 themſelves

-0.40

 deſt

-0.40

POSITIVE LOGITS

HomeAsUpEnabled

0.49

今度は

0.44

 Polsek

0.42

méable

0.41

SpringBootTest

0.41

addition

0.41

setupUi

0.41

expandindo

0.40

追加

0.40

Thêm

0.40

Activations Density 0.105%

phrases indicating addition or elaboration on previous points

another or also plus description

No Comments

No Known Activations