INDEX

Explanations

note

np_max-act · gemini-2.0-flash

The neuron fires on individual words that are dynamic action verbs or gerunds—especially high-energy, “do-something” words (e.g. nail, rock, roll, stomping, tackling, serving, run, score, cook up).

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

Negative Logits

_manager

-0.07

_deposit

-0.07

egasus

-0.06

 注意

-0.06

 Canal

-0.06

bows

-0.06

 Carpenter

-0.06

 Books

-0.06

-offset

-0.06

-around

-0.06

POSITIVE LOGITS

 ışı

0.07

هوری

0.07

чої

0.07

heed

0.06

천

0.06

 assists

0.06

(Str

0.06

({↵↵

0.06

 biking

0.06

 gpio

0.06

Activations Density 0.063%

note

The neuron fires on individual words that are dynamic action verbs or gerunds—especially high-energy, “do-something” words (e.g. nail, rock, roll, stomping, tackling, serving, run, score, cook up).

No Known Activations

note

The neuron fires on individual words that are dynamic action verbs or gerunds—especially high-energy, “do-something” words (e.g. nail, rock, roll, stomping, tackling, serving, run, score, cook up).

No Known Activations