INDEX

Explanations

numerical codes and values

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 36-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_36/width_16k/average_l0_120

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.36.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

+:+

-0.59

><><

-0.57

}</

-0.55

iſen

-0.54

mait

-0.52

 interloc

-0.50

alsy

-0.50

定向

-0.50

iſchen

-0.50

wiſe

-0.50

POSITIVE LOGITS

0.97

0.46

AppMethodBeat

0.46

Bronnen

0.43

Климаты

0.42

BufferException

0.40

ers

0.39

0.38

voorbeeld

0.38

 nueve

0.38

Activations Density 0.975%

numerical codes and values

No Comments

No Known Activations