INDEX

Explanations

specific Java application errors and exceptions related to input or software execution issues

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

technical terms 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

Tikang

-0.46

diyesi

-0.43

 Wikimedijinoj

-0.41

deutig

-0.38

 continúas

-0.38

 Isles

-0.36

wright

-0.36

 arşivlendi

-0.35

ELEM

-0.35

 Ecclesiastical

-0.34

POSITIVE LOGITS

 desconocido

0.63

Personendaten

0.57

kulum

0.54

 normalidad

0.48

 generado

0.48

 sufrido

0.48

 exportación

0.48

 cumplido

0.47

 LoginForm

0.46

 ocupado

0.46

Activations Density 0.007%

specific Java application errors and exceptions related to input or software execution issues

technical terms and identifiers

No Comments

No Known Activations