INDEX

Explanations

words and phrases indicating the presence of healthcare-related topics

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B @ 7-gemmascope-res-16k

Configuration

google/gemma-scope-9b-pt-res/layer_7/width_16k/average_l0_92

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

16,384

Data Type

float32

Hook Name

blocks.7.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 незавершена

-0.59

\{\\

-0.49

ArgsConstructor

-0.48

 Soluble

-0.47

trimmed

-0.47

shirt

-0.47

hock

-0.47

soluble

-0.47

suitable

-0.46

switching

-0.45

POSITIVE LOGITS

怎樣

0.45

 milliard

0.40

 autorytatywna

0.40

 nakalista

0.40

Зноскі

0.39

IUrlHelper

0.38

 TType

0.38

不变

0.38

 zkou

0.38

 miliardi

0.38

Activations Density 1.036%

words and phrases indicating the presence of healthcare-related topics

No Known Activations