INDEX

Explanations

phrases that indicate insight or understanding in a context related to research or analysis

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

 pinulongan

-0.62

انيف

-0.57

//});

-0.53

Abit

-0.53

✨:

-0.53

 subtotal

-0.53

әрмәләр

-0.52

 <<<<<<<<<<<<<<

-0.51

uxxxx

-0.51

extAlignment

-0.51

POSITIVE LOGITS

how

0.96

 what

0.84

the

0.74

why

0.69

hva

0.60

 cómo

0.59

what

0.57

 cuáles

0.54

 bagaimana

0.54

 πως

0.53

Activations Density 0.053%

phrases that indicate insight or understanding in a context related to research or analysis

No Comments

No Known Activations