INDEX

Explanations

collocations and phrases that indicate relationships or sequences among entities

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

New Auto-Interp

Top Features by Cosine Similarity

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

Configuration

google/gemma-scope-2b-pt-res/layer_7/width_16k/average_l0_69

Prompts (Dashboard)

36,864 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

Not in Any Lists

No Comments

Negative Logits

примеча

-0.50

ValueGeneration

-0.47

<caption>

-0.46

unsplash

-0.44

пус

-0.44

 inclusief

-0.43

察

-0.43

 Beteiligung

-0.43

FOLK

-0.43

ρίου

-0.42

POSITIVE LOGITS

 successive

0.86

 consecutive

0.74

 adjacent

0.72

 neighbouring

0.71

 adjoining

0.70

homonymie

0.68

 einander

0.68

 overlapping

0.68

 consecutively

0.68

曖昧さ回避

0.66

Activations Density 0.528%

collocations and phrases that indicate relationships or sequences among entities

No Comments

No Known Activations