INDEX

Explanations

utas

np_max-act · gemini-2.0-flash

The neuron selectively activates on personal proper names (e.g. first names of individuals).

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

esture

-0.07

plu

-0.06

ROUP

-0.06

 hebben

-0.06

,number

-0.06

iais

-0.06

dol

-0.06

ZIP

-0.06

 readOnly

-0.05

 chuẩn

-0.05

POSITIVE LOGITS

 Pascal

0.10

 Серг

0.08

řich

0.08

 Roberto

0.07

ří

0.07

 Petr

0.07

 Antonio

0.07

islav

0.07

datable

0.07

 Bram

0.07

Activations Density 0.074%

utas

The neuron selectively activates on personal proper names (e.g. first names of individuals).

No Comments

No Known Activations