INDEX

Explanations

specific references to various categories or classifications, particularly in a regulatory, demographic, or distinct grouping context

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.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

coroutines

-0.41

 continúas

-0.39

tanleria

-0.39

///</

-0.37

matchCondition

-0.36

 terdekat

-0.36

ViewInit

-0.36

Atsauces

-0.35

 Vordergrund

-0.35

 cortinas

-0.35

POSITIVE LOGITS

 ועל

0.70

 روی

0.70

onAdd

0.68

 وعلى

0.66

onSearch

0.65

 behalf

0.63

 platforms

0.61

on

0.60

onCancel

0.59

feinander

0.59

Activations Density 0.202%

specific references to various categories or classifications, particularly in a regulatory, demographic, or distinct grouping context

No Comments

No Known Activations