INDEX

Explanations

actions and events involving people or characters entering or exiting spaces

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 13-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_13/width_65k/average_l0_75

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.13.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

ALR

-0.56

 Acon

-0.52

 ivelany

-0.52

EndInit

-0.52

 Autorisations

-0.51

 بيها

-0.51

SAH

-0.49

ADM

-0.49

 <<<<<<<<<<<<<<

-0.48

SAI

-0.48

POSITIVE LOGITS

$__

0.68

ThemeOverlay

0.66

 soudain

0.62

 frappé

0.60

cyklopedia

0.57

hidupan

0.56

 bandeja

0.56

 tirer

0.54

 unexpectedly

0.54

 suddenly

0.54

Activations Density 0.027%

actions and events involving people or characters entering or exiting spaces

No Comments

No Known Activations