INDEX

Explanations

specific names or terms related to entities, particularly people and places

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-mlp/layer_6/width_131k/average_l0_72

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.6.hook_mlp_out

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

nodoc

-0.50

 ProtoMessage

-0.45

govina

-0.39

Derbyniad

-0.39

 Mind

-0.38

GEBURTSDATUM

-0.38

 "..\..\

-0.37

 Original

-0.37

master

-0.37

wijl

-0.37

POSITIVE LOGITS

queryInterface

0.47

сылкі

0.45

 interessiert

0.40

 fertil

0.39

 vermic

0.39

parsedMessage

0.38

 dessutom

0.37

 metast

0.36

 refusé

0.36

Билгалдахарш

0.36

Activations Density 0.013%

specific names or terms related to entities, particularly people and places

No Comments

No Known Activations