INDEX

Explanations

names of people and organizations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_16k/average_l0_68

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 itſelf

-0.48

intios

-0.47

 intermédiaire

-0.45

writeFieldEnd

-0.44

 način

-0.41

 ſmall

-0.41

 greateſt

-0.41

 themſelves

-0.41

addCriterion

-0.41

 beſt

-0.41

POSITIVE LOGITS

 betweenstory

0.49

ArrowToggle

0.48

('.',

0.45

仰

0.43

imedes

0.42

 Rockville

0.41

$^\

0.41

 Varma

0.41

 Réponses

0.41

 Ives

0.41

Activations Density 0.247%

names of people and organizations

No Comments

No Known Activations