INDEX

Explanations

code-related keywords and functions, particularly those pertaining to data structure operations and HTML element manipulations

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

keywords and identifiers

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

 <<<<<<<<<<<<<<

-0.52

 تضيفلها

-0.50

>",

-0.47

 disambiguazione

-0.45

+#+

-0.45

 betrekking

-0.44

Slf

-0.44

fastjson

-0.43

HttpFoundation

-0.42

 opgenomen

-0.42

POSITIVE LOGITS

sizePolicy

0.45

 Roskov

0.42

ंदीखरीदारी

0.40

ientôt

0.40

 Púb

0.39

 Gouvernement

0.39

WithValue

0.37

exitRule

0.35

FormTagHelper

0.35

 quim

0.35

Activations Density 0.236%

code-related keywords and functions, particularly those pertaining to data structure operations and HTML element manipulations

keywords and identifiers

No Comments

No Known Activations