INDEX

Explanations

references to specific scientific or technical datasets and analyses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 34-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_34/width_16k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.34.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

textra

-0.48

 bershka

-0.48

 famously

-0.47

 biodiversity

-0.47

timents

-0.47

 gonzález

-0.46

 cafés

-0.46

 stakeholders

-0.46

 datasets

-0.46

POSITIVE LOGITS

 mußten

0.56

 Schluß

0.54

faßt

0.51

 skall

0.48

luß

0.47

 ProtoMessage

0.46

 fué

0.46

 mußte

0.45

InSection

0.45

zeß

0.45

Activations Density 0.160%

references to specific scientific or technical datasets and analyses

No Comments

No Known Activations