INDEX

Explanations

academic references and formal proofs

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

addContainerGap

-0.53

 poveznice

-0.50

 fubject

-0.48

exitRule

-0.47

发表于

-0.45

addGap

-0.43

 itſelf

-0.43

altern

-0.42

TokenNameDOT

-0.41

rhy

-0.41

POSITIVE LOGITS

 autorytatywna

0.65

$_['

0.50

 مشين

0.49

ValueStyle

0.46

FormTagHelper

0.45

 للمعارف

0.42

TabIndex

0.41

 OFDb

0.41

 препратки

0.39

 zijne

0.39

Activations Density 0.356%

academic references and formal proofs

No Comments

No Known Activations