INDEX

Explanations

phrases that refer to levels and patterns within various contexts

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

the

-0.44

 tärke

-0.40

 many

-0.39

 venait

-0.39

 seura

-0.38

หลวง

-0.37

 gouver

-0.37

 portions

-0.36

 Bibliothek

-0.36

 sampah

-0.36

POSITIVE LOGITS

webElementXpaths

1.05

ngdoc

0.98

IUrlHelper

0.93

parsedMessage

0.90

 ujednoznacz

0.88

uxxxx

0.87

TypedDataSet

0.85

 Мексичка

0.83

ValueStyle

0.83

 kaarangay

0.83

Activations Density 0.792%

phrases that refer to levels and patterns within various contexts

No Comments

No Known Activations