INDEX

Explanations

code-related components or classes in programming environments

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

user turn starts

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

 StringField

-0.39

MarshalTo

-0.38

setVerticalGroup

-0.38

立てる

-0.37

Хьажоргаш

-0.36

 effusion

-0.35

 scan

-0.35

DeleteBehavior

-0.34

multirow

-0.34

fum

-0.34

POSITIVE LOGITS

Tembelea

0.60

 intptr

0.59

fromnode

0.52

 للاسماء

0.52

 kasarigan

0.50

 <<<<<<<<<<<<<<

0.49

endphp

0.48

 للمعارف

0.47

 betweenstory

0.47

uxxxx

0.45

Activations Density 0.001%

code-related components or classes in programming environments

user turn starts

No Comments

No Known Activations