INDEX

Explanations

elements related to function definitions and programming structures

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-16k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_16k/average_l0_91

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

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

tagHelperRunner

-0.80

bootstrapcdn

-0.76

Autoritní

-0.73

AddTagHelper

-0.65

 nahilalakip

-0.63

Tikang

-0.63

 صوتيه

-0.59

 Италијани

-0.59

delwed

-0.58

 surla

-0.58

POSITIVE LOGITS

 económica

0.48

 Económica

0.42

 anniversaire

0.41

 originated

0.40

 coucher

0.40

opsida

0.40

origin

0.40

PointerException

0.39

 itself

0.38

 convocatoria

0.37

Activations Density 0.713%

elements related to function definitions and programming structures

No Comments

No Known Activations