INDEX

Explanations

occurrences of the prefix "pla" and variations of the term "Hom."

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 2-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_2/width_16k/average_l0_67

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.2.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 <<<<<<<<<<<<<<

-0.67

 BoxFit

-0.47

はじめに

-0.47

gdala

-0.45

NewUrlParser

-0.45

 nák

-0.44

synthetic

-0.43

$_['

-0.43

 defaultstate

-0.42

 Nuss

-0.42

POSITIVE LOGITS

Pla

0.85

pla

0.83

Pla

0.75

pla

0.68

Mach

0.63

 Deleg

0.61

 barrel

0.60

deleg

0.60

 delegates

0.58

 deleg

0.58

Activations Density 0.603%

occurrences of the prefix "pla" and variations of the term "Hom."

No Comments

No Known Activations