INDEX

Explanations

references to titles, names, and honors associated with individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 21-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_21/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.21.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

➯

-0.70

########.

-0.68

tagHelperRunner

-0.58

 Lazio

-0.57

ResponseWriter

-0.57

 سكانية

-0.56

stdio

-0.56

 Drapeau

-0.56

 LoggerFactory

-0.55

Portale

-0.54

POSITIVE LOGITS

 mijne

0.73

 belangrij

0.63

 nôtre

0.59

 gevol

0.56

 dezelve

0.56

équi

0.54

vė

0.54

ulongan

0.52

cifix

0.52

 çik

0.51

Activations Density 1.923%

references to titles, names, and honors associated with individuals

No Comments

No Known Activations