INDEX

Explanations

references to shapes and their characteristics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 1-gemmascope-mlp-16k

Configuration

google/gemma-scope-2b-pt-mlp/layer_1/width_16k/average_l0_105

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.hook_mlp_out

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

 shape

-3.00

 Shape

-2.66

shape

-2.63

Shape

-2.42

 shapes

-2.39

 SHAPE

-2.11

shapes

-1.97

 Shapes

-1.96

SHAPE

-1.87

 shaped

-1.84

POSITIVE LOGITS

כום

0.48

 NSCoder

0.46

 folder

0.45

niad

0.44

хия

0.43

obel

0.42

oses

0.42

 Brin

0.42

 насељу

0.41

 clutch

0.41

Activations Density 0.171%

references to shapes and their characteristics

No Comments

No Known Activations