INDEX

Explanations

proper nouns, specifically names of individuals

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

+#+#

-0.70

 мәкал

-0.66

 Савезне

-0.63

hyrchwyd

-0.63

__(/*!

-0.58

 Chwiliwch

-0.58

 zufolge

-0.58

PostInfinity

-0.58

 للمعارف

-0.57

 potrivit

-0.57

POSITIVE LOGITS

 panik

0.32

 qualific

0.30

 calme

0.27

 Indon

0.27

SUMMARY

0.26

 inspección

0.26

 rense

0.25

Ayrıca

0.25

 incer

0.25

StoryboardSegue

0.25

Activations Density 0.024%

proper nouns, specifically names of individuals

No Comments

No Known Activations