INDEX

Explanations

keywords related to access modifiers and instance declarations in programming code

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

programming keywords

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_24/width_131k/average_l0_97

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.24.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

vs

-0.34

 results

-0.33

 come

-0.33

 llegado

-0.30

 always

-0.30

 when

-0.30

exercice

-0.30

 since

-0.30

了吗

-0.30

 seems

-0.29

POSITIVE LOGITS

ValueStyle

0.81

enumii

0.80

 Administrativna

0.77

rungsseite

0.74

таратура

0.73

 Normdatei

0.70

Билгалдахарш

0.67

posedge

0.67

CppMethod

0.66

fjspx

0.66

Activations Density 0.001%

keywords related to access modifiers and instance declarations in programming code

programming keywords

No Comments

No Known Activations