INDEX

Explanations

references to higher-order concepts or categorizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 4-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_4/width_16k/average_l0_124

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.4.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

üttel

-0.67

cul

-0.64

 Parton

-0.59

alah

-0.56

حوالہ

-0.56

 Alan

-0.56

-0.55

さ

-0.55

DateUtil

-0.54

åg

-0.54

POSITIVE LOGITS

 higher

2.19

 Higher

2.03

Higher

2.03

higher

2.02

 HIGHER

1.99

 höher

1.44

 höhere

1.35

 höheren

1.24

 lower

1.20

 högre

1.18

Activations Density 0.129%

references to higher-order concepts or categorizations

No Comments

No Known Activations