INDEX

Explanations

to

np_max-act · gemini-2.0-flash

This neuron activates on the word “According” (as in “According to …”), i.e. it detects attribution phrases introducing citations or sources.

oai_token-act-pair · o4-mini Triggered by @xinyanhu8

New Auto-Interp

Configuration

andyrdt/saes-llama-3.1-8b-instruct/resid_post_layer_11/trainer_1

Dataset (Dashboard)

Various

Features

131,072

Data Type

float32

Hook Name

blocks.11.hook_resid_post

Architecture

standard

Context Size

1,024

Dataset

monology/pile-uncopyrighted

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 مشار

-0.07

_daily

-0.07

 Lane

-0.07

pun

-0.07

 sublicense

-0.07

 Phone

-0.06

 shred

-0.06

Gallery

-0.06

лоб

-0.06

 Alpine

-0.06

POSITIVE LOGITS

告诉

0.07

According

0.07

 selon

0.07

 نتایج

0.07

Besides

0.07

 aseg

0.06

 secular

0.06

ρο

0.06

ідно

0.06

 según

0.06

Activations Density 0.018%

to

This neuron activates on the word “According” (as in “According to …”), i.e. it detects attribution phrases introducing citations or sources.

No Comments

No Known Activations