INDEX

Explanations

references to crashes and collisions

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

"Crash" or sounds suggesting impact

np_acts-logits-general · gemini-2.0-flash

crash course or impact

np_acts-logits-general · gemini-2.5-flash-lite

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 20-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_20/width_16k/average_l0_71

Prompts (Dashboard)

36,864 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

NOPQRST

-0.54

bpy

-0.51

 AspNetCore

-0.51

 fila

-0.48

 capelli

-0.48

orale

-0.48

énario

-0.48

ifera

-0.46

 dày

-0.46

gna

-0.45

POSITIVE LOGITS

 Monfieur

0.76

💥

0.70

 Cras

0.68

proof

0.66

DockStyle

0.66

 Reſ

0.65

 BOOM

0.65

 Diſ

0.65

SBATCH

0.65

 Jefus

0.65

Activations Density 0.254%

references to crashes and collisions

"Crash" or sounds suggesting impact

crash course or impact

No Comments

No Known Activations