INDEX

Explanations

critical terms and phrases indicating mathematical simplifications or conditions in theoretical discussions

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

 unarmed

-0.52

TTI

-0.48

MetaObject

-0.46

 crowdfunding

-0.41

 entertainers

-0.40

 movements

-0.40

 camping

-0.40

 superstitious

-0.40

SharedCtor

-0.40

 premed

-0.39

POSITIVE LOGITS

发表于

0.59

帖最后由

0.58

 idéia

0.53

 Bewußt

0.52

setVerticalGroup

0.51

脚注の使い方

0.49

 Aufwand

0.46

 avantages

0.46

 biß

0.45

 Schluß

0.45

Activations Density 0.456%

critical terms and phrases indicating mathematical simplifications or conditions in theoretical discussions

No Comments

No Known Activations