INDEX

Explanations

identifiers and structures related to programming code, especially within variable assignments and conditions

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

records, matches, Pub/Sub, user

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

__(/*!

-0.42

قایناق‌لار

-0.42

Erreferentziak

-0.39

 dolci

-0.38

 Leder

-0.38

 ***!

-0.37

elve

-0.36

当て

-0.36

 PeEnEo

-0.35

komme

-0.35

POSITIVE LOGITS

 surla

0.49

MessageTagHelper

0.45

 تضيفلها

0.43

Captor

0.42

 afterwards

0.41

verifyException

0.39

เหน

0.39

Emerging

0.36

 afterward

0.36

AutoField

0.36

Activations Density 0.133%

identifiers and structures related to programming code, especially within variable assignments and conditions

records, matches, Pub/Sub, user

No Comments

No Known Activations