INDEX

Explanations

named entities and organizations within various contexts

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

organizations and locations

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

mentaux

-0.33

حياتها

-0.32

 tqdm

-0.30

üstü

-0.29

 ویکی‌پدی

-0.28

 obicei

-0.28

イクアウト

-0.28

 full

-0.28

 natürliche

-0.28

 otomatig

-0.27

POSITIVE LOGITS

ValueStyle

0.69

 utafitiHapana

0.63

 ModelExpression

0.63

ConstraintMaker

0.60

الدراسه

0.59

prepareStatement

0.58

########.

0.57

URLException

0.55

RegressionTest

0.55

-------

0.53

Activations Density 0.129%

named entities and organizations within various contexts

organizations and locations

No Comments

No Known Activations