INDEX

Explanations

|--

np_max-act · gemini-2.0-flash

This neuron detects ASCII‐art directory‐tree drawing characters (e.g. the ├──, └──, │, ── tokens used to show folder structures).

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

file and directory structures within programming projects.

oai_token-act-pair · gpt-4o-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

 нього

-0.07

 Lena

-0.07

terní

-0.07

 них

-0.07

ber

-0.07

 удар

-0.07

形式

-0.07

 Device

-0.07

ailles

-0.06

 зна

-0.06

POSITIVE LOGITS

contra

0.06

_input

0.06

 Amen

0.06

达到

0.06

}-{

0.06

ـــ

0.06

oğraf

0.06

 trespass

0.06

max

0.05

Activations Density 0.003%

|--

This neuron detects ASCII‐art directory‐tree drawing characters (e.g. the ├──, └──, │, ── tokens used to show folder structures).

file and directory structures within programming projects.

No Comments

No Known Activations