INDEX

Explanations

phrases indicating preferences or choices, particularly in favor of alternatives and substitutive actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

完

-0.36

bigsqcup

-0.35

fil

-0.35

 conci

-0.35

 context

-0.33

contextMenu

-0.32

 CORON

-0.32

 saber

-0.31

 partic

-0.31

 antig

-0.30

POSITIVE LOGITS

tdessen

0.80

verwijspagina

0.75

MLLoader

0.75

StoryboardSegue

0.69

Instead

0.68

代わりに

0.65

 Instead

0.64

InputTagHelper

0.63

 instead

0.63

FXMLLoader

0.63

Activations Density 0.604%

phrases indicating preferences or choices, particularly in favor of alternatives and substitutive actions

No Comments

No Known Activations