INDEX

Explanations

references to data validation and presence checks in a coding or legal context

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

"Warden" in legal contexts

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_32/width_131k/average_l0_88

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.32.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

rungsseite

-0.76

المشاركات

-0.54

Билгалдахарш

-0.52

ervlak

-0.52

 ujednoznacz

-0.51

parsedMessage

-0.49

Földrajzportál

-0.49

dyž

-0.48

 anún

-0.48

 miniaturka

-0.47

POSITIVE LOGITS

fjspx

1.05

oucí

0.68

"}")

0.50

jury

0.41

Rom

0.40

printStackTrace

0.40

DebuggerNonUser

0.40

Geplaatst

0.40

 AspNetCore

0.39

...).

0.39

Activations Density 0.000%

references to data validation and presence checks in a coding or legal context

"Warden" in legal contexts

No Comments

No Known Activations