INDEX

Explanations

phrases related to causation and errors

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 11-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_11/width_16k/average_l0_80

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.11.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

 reindeer

-0.60

 >=",

-0.55

 Meksiku

-0.52

 Italijanski

-0.52

 masturb

-0.51

 Waray

-0.50

ovi

-0.50

 Kure

-0.50

TypedDataSet

-0.49

 Mazar

-0.49

POSITIVE LOGITS

tagext

0.67

Viungo

0.63

 inkább

0.61

Kanpo

0.60

$_['

0.60

amaño

0.60

reportWebVitals

0.56

むしろ

0.55

 okuyayım

0.54

ásban

0.54

Activations Density 0.418%

phrases related to causation and errors

No Comments

No Known Activations