INDEX

Explanations

descriptors related to quality and reputation in film production

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

 ſtate

-0.52

Rujuakan

-0.51

 كومونز

-0.49

 houſe

-0.48

 ſind

-0.46

 ſche

-0.44

 purpoſe

-0.44

 Waray

-0.43

 queſta

-0.41

 Houſe

-0.41

POSITIVE LOGITS

 reír

0.51

 gewohnt

0.43

 lyrical

0.43

 trademark

0.43

 Cuevas

0.41

Grüsse

0.41

 Geruch

0.40

FlatAppearance

0.40

PYX

0.40

 Kindern

0.39

Activations Density 0.028%

descriptors related to quality and reputation in film production

No Comments

No Known Activations