INDEX

Explanations

references to the Roman Church or Roman-related subjects

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 1-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_1/width_16k/average_l0_69

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

 pill

-0.40

coe

-0.38

 ско

-0.37

 Nebel

-0.36

DNS

-0.36

 referrerpolicy

-0.36

 crusher

-0.36

printer

-0.35

 pedicure

-0.35

ndice

-0.35

POSITIVE LOGITS

rom

1.01

RM

0.98

Rom

0.97

rm

0.96

Rom

0.96

 roman

0.90

Rm

0.88

ROM

0.88

 Roman

0.85

Rm

0.85

Activations Density 0.297%

references to the Roman Church or Roman-related subjects

No Comments

No Known Activations