INDEX

Explanations

Probability letter picking questions

np_max-act · gemini-2.0-flash

This neuron detects tokens of the form “<letter>:<number>,” i.e. letter-colon-numeric entries in the multiset weight/probability lists.

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

 Dmit

-0.06

trie

-0.06

 тка

-0.06

Leg

-0.06

 уг

-0.06

dbe

-0.06

PMC

-0.06

.MATCH

-0.06

 mest

-0.06

hoe

-0.06

POSITIVE LOGITS

 totalitarian

0.07

 sevent

0.06

.*,

0.06

 topl

0.06

 내려

0.06

 нарез

0.06

_Per

0.06

 Advocate

0.06

 COMMAND

0.06

 culpa

0.06

Activations Density 0.003%

Probability letter picking questions

This neuron detects tokens of the form “<letter>:<number>,” i.e. letter-colon-numeric entries in the multiset weight/probability lists.

No Comments

No Known Activations