INDEX

Explanations

phrases that indicate harm or damage to objects or entities

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-2b-pt-res/layer_9/width_16k/average_l0_73

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.9.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

 referenties

-0.82

roxene

-0.65

 Boyer

-0.64

 hujan

-0.63

Underline

-0.61

edoc

-0.60

 autorytatywna

-0.59

 Fowler

-0.58

 Finley

-0.57

Према

-0.57

POSITIVE LOGITS

 damage

1.74

 damages

1.71

 DAMAGE

1.65

 Damage

1.64

 Damages

1.63

Damage

1.53

damage

1.53

 DAMAGES

1.46

 Damaged

1.38

 damaged

1.37

Activations Density 0.087%

phrases that indicate harm or damage to objects or entities

No Comments

No Known Activations