INDEX

Explanations

references to specialized units or components 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-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

pompa

-0.42

 канали

-0.41

pelier

-0.40

ylen

-0.39

 Spel

-0.38

fromnode

-0.38

chin

-0.38

OGND

-0.38

たまた

-0.37

 wasn

-0.37

POSITIVE LOGITS

 nemlig

0.58

兩種

0.56

Jereo

0.56

RTLR

0.54

ParallelGroup

0.51

новништво

0.47

ariConfig

0.46

المشاركات

0.45

两种

0.45

ArrowToggle

0.44

Activations Density 0.124%

references to specialized units or components within various contexts

No Comments

No Known Activations