INDEX

Explanations

references to various 'programmes' or 'programs'

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-res-16k

Configuration

google/gemma-scope-2b-pt-res/layer_1/width_16k/average_l0_102

Prompts (Dashboard)

36,864 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.1.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

 costos

-0.58

 tumors

-0.54

tis

-0.52

ing

-0.52

 bris

-0.52

iting

-0.52

ism

-0.52

cing

-0.51

 fibers

-0.50

سه

-0.50

POSITIVE LOGITS

 programme

2.57

Programme

2.46

 Programme

2.42

programme

2.38

 programmes

2.30

 PROGRAMME

2.27

 Programmes

2.00

 Programm

1.00

 grammes

0.99

sprogramm

0.96

Activations Density 0.078%

references to various 'programmes' or 'programs'

No Comments

No Known Activations