INDEX

Explanations

terms related to design specifications and intended functions of various programs and tools

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

Negative Logits

министра

-0.35

 possibilités

-0.34

 suivants

-0.33

 endblock

-0.31

ổi

-0.31

abestanden

-0.30

 indé

-0.30

HasAnnotation

-0.30

ーズ

-0.29

Charsets

-0.29

POSITIVE LOGITS

 designed

2.50

 intended

2.34

designed

2.20

 meant

2.13

intended

2.11

 Designed

2.09

 Intended

2.06

Designed

1.99

meant

1.93

Intended

1.84

Activations Density 0.868%

terms related to design specifications and intended functions of various programs and tools

No Known Activations