INDEX

Explanations

coding constructs related to method definitions and function overrides

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

Java method names

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

 يتيمه

-0.74

Parcelize

-0.60

⋙

-0.53

Slf

-0.52

 Tsch

-0.52

 rootReducer

-0.50

\{\\

-0.50

XOR

-0.50

btnClose

-0.48

 ProgressDialog

-0.48

POSITIVE LOGITS

AdapterView

2.70

 AdapterView

1.85

BaseAdapter

0.68

setOnItem

0.64

CompoundButton

0.64

ChildScrollView

0.60

 argint

0.44

 ArrayAdapter

0.43

 препратки

0.43

 lentejuelas

0.43

Activations Density 0.000%

coding constructs related to method definitions and function overrides

Java method names

No Comments

No Known Activations