INDEX

Explanations

references to specific individuals and their actions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 38-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_38/width_16k/average_l0_128

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.38.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

 agré

-0.38

 Expédié

-0.35

 docentes

-0.35

 étend

-0.35

valg

-0.33

Manufacturing

-0.33

tuple

-0.33

 Wisata

-0.33

oliber

-0.33

 spazi

-0.33

POSITIVE LOGITS

 wrestling

1.08

 wrestled

1.03

 wrestlers

0.99

 wrestle

0.98

 wrestler

0.97

 Wrestling

0.94

AEW

0.90

Wrestling

0.83

 wrest

0.80

 Wrestle

0.80

Activations Density 0.085%

references to specific individuals and their actions

No Comments

No Known Activations