INDEX

Explanations

programming-related commands or keywords

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

Personendaten

-0.43

 esclavos

-0.41

canestro

-0.40

eseorang

-0.38

 oídos

-0.37

 dépens

-0.37

omock

-0.37

 referirse

-0.35

 Gemü

-0.35

aperto

-0.34

POSITIVE LOGITS

andExpect

0.73

SourceChecksum

0.68

ACHUSET

0.65

MIDDLEWARE

0.63



0.61

Hentet

0.57

 migra

0.57

 Référence

0.55

IndentedString

0.53

$_"

0.52

Activations Density 0.000%

programming-related commands or keywords

No Comments

No Known Activations