INDEX

Explanations

programming language syntax and structure related to conditional statements and object property checks

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

then

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

ValueStyle

-0.65

ItemBackground

-0.56

expandindo

-0.55

IntoConstraints

-0.55

 arşivlendi

-0.55

ftagPool

-0.54

InputTagHelper

-0.53

DIPSETTING

-0.51

AccessorTable

-0.51

 okuyayım

-0.50

POSITIVE LOGITS

kapet

0.35

 latter

0.34

 apellidos

0.33

recognize

0.33

aperto

0.32

tenis

0.32

 Saito

0.32

 defaultstate

0.32

testcase

0.32

iança

0.32

Activations Density 0.023%

programming language syntax and structure related to conditional statements and object property checks

then

No Comments

No Known Activations