INDEX

Explanations

ka

np_max-act · gemini-2.0-flash

This neuron detects non-English (Cyrillic‐script) tokens—i.e. snippets of text in Russian (and similar) rather than in Latin-script English.

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

We

-0.10

We

-0.09

we

-0.09

↵

-0.07

us

-0.07

aptop

-0.07

↵↵↵↵↵↵↵↵

-0.06

ニﾆ

-0.06

our

-0.06

，因此

-0.06

POSITIVE LOGITS

 Yellow

0.06

 their

0.06

 they

0.06

his

0.06

createElement

0.06

/docs

0.06

rub

0.06

 підтрим

0.06

 yellow

0.06

 World

0.06

Activations Density 0.205%

ka

This neuron detects non-English (Cyrillic‐script) tokens—i.e. snippets of text in Russian (and similar) rather than in Latin-script English.

No Comments

No Known Activations