INDEX

Explanations

prepositions and indicators of location or context

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

 laſſen

-0.85

 Waſſer

-0.78

ésultats

-0.77

 queſta

-0.77

 RIPRODUZIONE

-0.77

 verſch

-0.77

 müſſen

-0.77

 beſch

-0.76

 wiſſen

-0.76

 témoig

-0.75

POSITIVE LOGITS

in

1.13

In

1.12

In

1.09

in

1.09

IN

1.05

IN

1.00

isIn

0.82

getIn

0.74

 isIn

0.67

в

0.67

Activations Density 0.074%

prepositions and indicators of location or context

No Comments

No Known Activations