INDEX

Explanations

references to training and educational activities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 37-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_37/width_16k/average_l0_124

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.37.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

 dieux

-0.46

 TMPro

-0.45

 vandens

-0.44

 atve

-0.44

huriyet

-0.42

 fieltro

-0.42

ufficio

-0.42

degna

-0.42

 comerciais

-0.41

itectura

-0.41

POSITIVE LOGITS

 wheels

0.59

 sessions

0.55

 camps

0.54

 camp

0.52

 regimen

0.50

 regimens

0.50

lager

0.48

wheels

0.47

wreck

0.46

 regime

0.45

Activations Density 0.129%

references to training and educational activities

No Comments

No Known Activations