INDEX

Explanations

life experiences

np_max-act · gemini-2.0-flash

This neuron primarily responds to unusually long tokens (multisyllabic or multi‐character words), with activation strength roughly increasing as token length/complexity increases.

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

ánt

-0.07

reinterpret

-0.07

 thận

-0.07

_SF

-0.06

ifth

-0.06

 waiter

-0.06

 overall

-0.06

Traditional

-0.06

fmt

-0.06

Cin

-0.06

POSITIVE LOGITS

Helmet

0.06

=email

0.06

.setItems

0.06

 Coordinates

0.06

 stato

0.06

olkata

0.06

BOR

0.06

?");↵

0.06

DEN

0.06

↑

0.06

Activations Density 0.202%

life experiences

This neuron primarily responds to unusually long tokens (multisyllabic or multi‐character words), with activation strength roughly increasing as token length/complexity increases.

No Comments

No Known Activations