INDEX

Explanations

scientific and numerical expressions within text

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

Negative Logits

che

-0.34

jScrollPane

-0.32

否

-0.32

<?

-0.31

 اغ

-0.31

 Untersch

-0.31

tab

-0.30

handleClose

-0.30

ycznego

-0.29

POSITIVE LOGITS

verwijspagina

0.87

 myſelf

0.74

 ujednoznacz

0.73

expandindo

0.67

KommentareTeilen

0.65

Tikang

0.64

ſelf

0.60

ſelves

0.59

GTCX

0.59

 dortmund

0.59

Activations Density 0.238%

scientific and numerical expressions within text

No Known Activations