INDEX

Explanations

references to a specific group of people or entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

these

-0.37

Obrázky

-0.35

<bos>

-0.34

 these

-0.34

PRWEB

-0.34

ing

-0.34

gether

-0.32

simp

-0.29

and

-0.29

 Familienname

-0.29

POSITIVE LOGITS

 fleste

0.82

jenigen

0.79

Personendaten

0.77

 flesta

0.74

 involved

0.72

 pesky

0.69

 wenigen

0.68

zelfde

0.68

 betrokken

0.65

 betweenstory

0.65

Activations Density 0.125%

references to a specific group of people or entities

No Comments

No Known Activations