INDEX

Explanations

identifiers and classes related to data structures and functions within programming contexts

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

resource ID identifiers

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

 שוליים

-0.67

Hozzáférés

-0.57

 BoxDecoration

-0.57

ſſung

-0.57

 CURIAM

-0.56

஦

-0.56

nonatomic

-0.55

dataclass

-0.55

<unused58>

-0.55

<unused15>

-0.55

POSITIVE LOGITS

ID

0.42

Personensuche

0.40

code

0.38

ID

0.38

id

0.38

idUser

0.37

id

0.36

Id

0.35

Id

0.32

 незавершена

0.32

Activations Density 0.004%

identifiers and classes related to data structures and functions within programming contexts

resource ID identifiers

No Comments

No Known Activations