INDEX

Explanations

references to programming constructs or entities in code

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

foreign words, symbols

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

 '\\;'

-0.69

 pinulongan

-0.65

HtmlAttribute

-0.55

 betweenstory

-0.54

ChildScrollView

-0.54

nicity

-0.54

 Lenn

-0.53



-0.52

feit

-0.52

chuhe

-0.52

POSITIVE LOGITS

 obyvateľov

0.37

 eût

0.36

 виправивши

0.35

 frumo

0.35

 frumos

0.34

 hvit

0.34

 mourir

0.34

enumii

0.34

 informée

0.33

enumi

0.32

Activations Density 0.004%

references to programming constructs or entities in code

foreign words, symbols

No Comments

No Known Activations