INDEX

Explanations

references to in-person interactions or interviews

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

 InputDecoration

-0.57

 informée

-0.55

SafeMath

-0.54

pecabe

-0.52

stylers

-0.48

 Wikimedijinoj

-0.48

 gespannt

-0.48

Gweler

-0.47

 AssemblyCulture

-0.47

 wireType

-0.46

POSITIVE LOGITS

person

0.46

 presencial

0.44

PC

0.42

nteral

0.41

INCREF

0.40

physical

0.40

 meetings

0.40

pc

0.40

intptr

0.39

AspNetCore

0.39

Activations Density 0.007%

references to in-person interactions or interviews

No Comments

No Known Activations