INDEX

Explanations

references to specific technical elements or identifiers in a detailed context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-9B-IT @ 20-gemmascope-res-131k

Configuration

google/gemma-scope-9b-it-res/layer_20/width_131k/average_l0_81

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

 Annahme

-0.47

 spreken

-0.42

 vergeten

-0.41

 ferien

-0.41

tanleria

-0.41

autaire

-0.39

 çalışan

-0.38

 verheiratet

-0.38

principalColumn

-0.37

 Folgen

-0.37

POSITIVE LOGITS

Autoritní

0.58

 autorytatywna

0.55

</tfoot>

0.50

GEBURTSDATUM

0.47

toprule

0.47

')):

0.46

 Signalez

0.46

 ostavi

0.45

"):

0.45

uxxxx

0.45

Activations Density 0.324%

references to specific technical elements or identifiers in a detailed context

No Comments

No Known Activations