INDEX

Explanations

references to "LG" and terms related to singular or unique entities

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GEMMA-2-2B @ 0-gemmascope-res-65k

Configuration

google/gemma-scope-2b-pt-res/layer_0/width_65k/average_l0_73

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

monology/pile-uncopyrighted

Features

65,536

Data Type

float32

Hook Name

blocks.0.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 chancel

-0.62

 rigor

-0.58

 morire

-0.56

 menikah

-0.55

 inoxydable

-0.55

 subroutine

-0.55

 scattata

-0.55

 Gret

-0.54

EOF

-0.54

 vorrei

-0.54

POSITIVE LOGITS

aarrggbb

0.70

LG

0.67

AddTagHelper

0.66

openqa

0.65

vaux

0.65

'{@

0.63

parsedMessage

0.61

lg

0.59

 utafitiHapana

0.59

 lonely

0.58

Activations Density 0.076%

references to "LG" and terms related to singular or unique entities

No Comments

No Known Activations