INDEX

Explanations

references to conflict and targeted actions against individuals in a sociopolitical context

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

 فريبيس

-0.40

 lenker

-0.39

曖昧さ回避

-0.39

GrantedAuthority

-0.38

 óptica

-0.38

 komma

-0.37

tahui

-0.36

RunWith

-0.36

cientes

-0.36

 händer

-0.35

POSITIVE LOGITS

//

0.51

 rival

0.50

 enemies

0.50

MethodManager

0.49

 rivals

0.47

 targets

0.47

/*

0.46

GTCX

0.44

 protoimpl

0.43

 TARGET

0.42

Activations Density 0.525%

references to conflict and targeted actions against individuals in a sociopolitical context

No Comments

No Known Activations