INDEX

Explanations

numerical dimensions and specifications of objects

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

 étrangères

-0.52

 étrangers

-0.49

ptonshire

-0.46

 étranger

-0.45

 étrangère

-0.44

tifice

-0.42

jectures

-0.41

üte

-0.41

 DEFENSE

-0.40

WriteBarrier

-0.40

POSITIVE LOGITS

 متحده

0.38

 AssemblyCulture

0.38

tafel

0.37

Mitte

0.35

 doors

0.35

 console

0.34

carre

0.34

 Frie

0.33

 doInBackground

0.33

furn

0.33

Activations Density 0.002%

numerical dimensions and specifications of objects

No Comments

No Known Activations