INDEX

Explanations

technical terms related to programming or coding syntax

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

Java code mnemonics and foreign characters

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

 surla

-0.66

 zijne

-0.65

 protoimpl

-0.61

Datuak

-0.60

 itſelf

-0.58

 Anſ

-0.57

حياته

-0.56

 ucapnya

-0.56

 تعدى

-0.56

+#+#

-0.55

POSITIVE LOGITS

 onOptions

0.57

RTEE

0.44

Carriera

0.43

 visa

0.37

LookAnd

0.36

 BorderLayout

0.35

'{@

0.35

 titular

0.35

 lungo

0.35

 quid

0.35

Activations Density 0.000%

technical terms related to programming or coding syntax

Java code mnemonics and foreign characters

No Comments

No Known Activations