INDEX

Explanations

terms related to fragments and processes in scientific texts

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.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

########.

-0.64

UrlResolution

-0.63

UnusedPrivate

-0.55

XmlEnum

-0.54

 дописавши

-0.52

 betweenstory

-0.52

adə

-0.52

AsUp

-0.50

✨:

-0.50

 ModelExpression

-0.49

POSITIVE LOGITS

 sprite

0.72

sprite

0.69

 Sprite

0.68

Spr

0.68

SPR

0.67

 spray

0.67

Sprite

0.67

 spraying

0.67

 sprayed

0.64

 Spray

0.60

Activations Density 0.385%

terms related to fragments and processes in scientific texts

No Comments

No Known Activations