INDEX

Explanations

Questions

np_max-act · gemini-2.0-flash

The neuron selectively lights up on numeric tokens and digit‐heavy subwords (e.g. version numbers, file paths, decimal numbers).

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

Selection

-0.07

ional

-0.06

}}],↵

-0.06

 economists

-0.06

:t

-0.06

<Item

-0.06

 Ionic

-0.06

 Protestant

-0.06

_extra

-0.06

?>"/>↵

-0.06

POSITIVE LOGITS

regn

0.07

近

0.06

 blurry

0.06

hy

0.06

 Flem

0.06

fun

0.06

aryana

0.06

 IDisposable

0.06

iff

0.06

 midi

0.06

Activations Density 0.000%

Questions

The neuron selectively lights up on numeric tokens and digit‐heavy subwords (e.g. version numbers, file paths, decimal numbers).

No Comments

No Known Activations

Questions

The neuron selectively lights up on numeric tokens and digit‐heavy subwords (e.g. version numbers, file paths, decimal numbers).

No Comments

No Known Activations