INDEX

Explanations

programming languages and frameworks

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

python coding contexts

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

лтемелер

-0.44

 například

-0.44

 späteren

-0.43

 privées

-0.41

 religieuses

-0.40

 esimerkiksi

-0.40

 dieselbe

-0.38

 derselben

-0.38

selben

-0.38

 dieselben

-0.37

POSITIVE LOGITS

 XCTest

0.58

 pageContext

0.57

AutoScale

0.56

 Coat

0.53

setVerticalGroup

0.52

RenderAtEndOf

0.51

 Reson

0.50

 plead

0.50

DockStyle

0.49

 coat

0.49

Activations Density 0.001%

programming languages and frameworks

python coding contexts

No Comments

No Known Activations