INDEX

Explanations

patterns or features related to technical specifications or framework components

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

af

-2.09

af

-2.03

AF

-1.78

AF

-1.77

Af

-1.74

Af

-1.63

afs

-1.02

afu

-0.93

afy

-0.91

afl

-0.90

POSITIVE LOGITS

 perfetto

0.54

 remercier

0.53

 españa

0.48

IALLY

0.48

 vrijwilli

0.47

Fech

0.47

като

0.47

 maxHeight

0.46

 volont

0.45

 enví

0.45

Activations Density 0.214%

patterns or features related to technical specifications or framework components

No Comments

No Known Activations