INDEX

Explanations

words or phrases that suggest a summary or introduction

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

 للاسماء

-0.71

 برانيه

-0.66

 nahilalakip

-0.63

BeginContext

-0.62

出版年

-0.60

 насељу

-0.60

OGND

-0.60

verwijspagina

-0.59

httphttps

-0.59



-0.57

POSITIVE LOGITS

 thermique

0.40

 köz

0.35

 complete

0.34

 Aprend

0.34

 INSTRUCTION

0.33

 Overall

0.33

Instru

0.32

 Instructions

0.32

 completo

0.31

 göre

0.31

Activations Density 0.000%

words or phrases that suggest a summary or introduction

No Comments

No Known Activations