INDEX

Explanations

steps and sequences in procedural content

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 vermelha

-0.45

 onBind

-0.43

 luzes

-0.42

 prolific

-0.40

 Harrington

-0.40

เย็น

-0.40

 kaik

-0.39

ugc

-0.38

แดง

-0.38

 refrigerador

-0.38

POSITIVE LOGITS

Step

1.34

 Step

1.30

 step

1.28

 Steps

1.28

Steps

1.26

 STEP

1.25

 steps

1.23

STEP

1.20

step

1.20

steps

1.16

Activations Density 0.073%

steps and sequences in procedural content

No Comments

No Known Activations