INDEX

Explanations

references to GPU capabilities and CUDA programming attributes

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

CUDA errors or device attributes

np_acts-logits-general · gemini-2.0-flash

CUDA and TensorFlow

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 22-gemmascope-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_22/width_16k/average_l0_72

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.22.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

warma

-0.49

 panel

-0.46

jPanel

-0.46

 Coleoptera

-0.46

iteritems

-0.45

NamedQueries

-0.44

δρο

-0.44

Kef

-0.44

еро

-0.43

 Combine

-0.43

POSITIVE LOGITS

 torch

1.16

torch

1.03

 torches

0.83

 Torch

0.80

 MainAxisSize

0.79

Torch

0.79

uxxxx

0.71

:✨

0.71

 torchvision

0.66

 tensor

0.61

Activations Density 0.083%

references to GPU capabilities and CUDA programming attributes

CUDA errors or device attributes

CUDA and TensorFlow

No Comments

No Known Activations