INDEX

Explanations

specific programming or technical terms related to data structures and manipulation

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

 Administrativna

-0.61

 disambiguazione

-0.55

 Valence

-0.51

 useContext

-0.50

aarrggbb

-0.49

 statistical

-0.49

iNdEx

-0.48

ifflin

-0.48

Rujuakan

-0.47

 unknownFields

-0.47

POSITIVE LOGITS

новништво

0.61

TagMode

0.44

webElementXpaths

0.41

ValueGenerated

0.39

principalTable

0.38

ArrowToggle

0.35

CloseOperation

0.33

richTextPanel

0.32

enumii

0.30

construct

0.30

Activations Density 0.033%

specific programming or technical terms related to data structures and manipulation

No Comments

No Known Activations