INDEX

Explanations

references to political campaigns and endorsements

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

 صوتيه

-0.56

 procu

-0.42

 Procu

-0.36

 quæ

-0.36

 world

-0.35

 desco

-0.32

Jereo

-0.32

 desem

-0.32

 racional

-0.31

ServiceClient

-0.31

POSITIVE LOGITS

setVerticalGroup

0.61

 campaigned

0.57

$_['

0.57

 campaigning

0.57

paign

0.55

 Democrat

0.53

MLLoader

0.53

 incumb

0.53

ViewImports

0.52

 candidacy

0.52

Activations Density 0.108%

references to political campaigns and endorsements

No Comments

No Known Activations