INDEX

Explanations

references to locations and situations involving harm or danger

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 0-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_0/width_16k/average_l0_129

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.0.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

 prêtre

-0.50

 enfans

-0.48

Demografia

-0.47

 downright

-0.47

 extranjero

-0.47

 diable

-0.47

 Sénat

-0.47

قایناقلار

-0.46

 Mexique

-0.46

 Trenton

-0.46

POSITIVE LOGITS

 harmony

0.69

 harmonic

0.66

 Harmony

0.63

 harmon

0.62

 Harbour

0.61

Harmony

0.59

andExpect

0.59

 Harmonic

0.59

LookAnd

0.56

verwijspagina

0.56

Activations Density 0.228%

references to locations and situations involving harm or danger

No Comments

No Known Activations