INDEX

Explanations

expressions related to sufficiency or adequacy

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 zonne

-0.37

-0.35

 Cardinale

-0.34

 Luar

-0.34

före

-0.32

↵↵

-0.31

 asistir

-0.30

 globally

-0.30

 pierdas

-0.30

POSITIVE LOGITS

Enough

1.52

 Enough

1.52

enough

1.48

 enough

1.45

 ENOUGH

1.37

Sufficient

1.31

 Sufficient

1.28

 sufficient

1.24

sufficient

1.20

 suffic

1.13

Activations Density 0.069%

expressions related to sufficiency or adequacy

No Comments

No Known Activations