INDEX

Explanations

phrases indicating approximate numerical quantities or time frames

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

New Auto-Interp

Configuration

google/gemma-scope-9b-pt-res/layer_20/width_131k/average_l0_114

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

131,072

Data Type

float32

Hook Name

blocks.20.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 transfieras

-0.57

 bezeichneter

-0.55

WebElementEntity

-0.52

⎨

-0.50

 Signalez

-0.48

 noDo

-0.46

 ویکی‌پدی

-0.44

 Infórmanos

-0.41

astify

-0.40

Hochspringen

-0.39

POSITIVE LOGITS

Portail

0.54

Архівовано

0.53

Something

0.52

쯤

0.52

 tertentu

0.51

Elsewhere

0.50

 ungefähr

0.50

こち

0.50

 Vielleicht

0.49

 wohl

0.48

Activations Density 0.007%

phrases indicating approximate numerical quantities or time frames

No Comments

No Known Activations