INDEX

Explanations

references to specific measurements, entities, or items related to micro-level details

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

 Economía

-0.40

 prêtre

-0.40

 contenu

-0.38

 cuenca

-0.36

 inauguración

-0.35

 gardien

-0.35

 GenerationType

-0.35

 CreateTagHelper

-0.35

vœ

-0.35

 Construcción

-0.34

POSITIVE LOGITS

:✨

0.66

RegistryLite

0.60

⟭

0.57

Derbyniad

0.57

انيف

0.57

 sumpay

0.54

oa̍t

0.53

 Roskov

0.52

RTDA

0.52

хьтан

0.52

Activations Density 0.205%

references to specific measurements, entities, or items related to micro-level details

No Comments

No Known Activations