INDEX

Explanations

references to origin or source of information

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

 perfección

-0.43

 sosok

-0.42

 normalidad

-0.37

 rehabilitación

-0.37

GEBURTSDATUM

-0.36

 secado

-0.36

 frågan

-0.36

 suministro

-0.36

 productividad

-0.36

 densidad

-0.35

POSITIVE LOGITS

RectangleBorder

0.69

 neighboring

0.64

CppMethod

0.60

usercontent

0.60

 South

0.58

 Southern

0.57

來自

0.57

بوابة

0.56

 nationalities

0.56

viewWillAppear

0.55

Activations Density 0.034%

references to origin or source of information

No Comments

No Known Activations