INDEX

Explanations

phrases that indicate causation or justifications

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 19-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_19/width_16k/average_l0_109

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.19.hook_mlp_out

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

 kasarigan

-0.52

 places

-0.51

MessageTagHelper

-0.49

 lĩnh

-0.47

while

-0.47

places

-0.47

spesies

-0.47

 clarity

-0.46

 pois

-0.46

 Clarity

-0.46

POSITIVE LOGITS

 для

0.75

for

0.74

存于互联网档案馆

0.69

 untuk

0.68

Diweddarwch

0.68

BeginContext

0.68

 để

0.62

для

0.61

สำหรับ

0.60

 للمعارف

0.60

Activations Density 0.313%

phrases that indicate causation or justifications

No Comments

No Known Activations