INDEX

Explanations

sequences of numerical values or identifiers

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-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

 Barbier

-0.49

//

-0.48

 horário

-0.45

 calmer

-0.44

neurial

-0.43

 aquarela

-0.42

 suitable

-0.41

endsection

-0.41

 palettes

-0.41

Mutability

-0.41

POSITIVE LOGITS

tagHelperRunner

0.77

0.60

 Tenth

0.52

tenth

0.50

HomeAsUpEnabled

0.48

 kaarangay

0.48

 quæ

0.47

OrWhiteSpace

0.46

ldc

0.45

 tenth

0.45

Activations Density 0.026%

sequences of numerical values or identifiers

No Comments

No Known Activations