INDEX

Explanations

characters

np_max-act · gemini-2.0-flash

The neuron activates on meta‐narrative tokens—words that refer to the fictional framing or characters (e.g. “character,” “characters,” “fiction,” “scenario”).

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

τηκε

-0.07

(surface

-0.07

ılığ

-0.06

acija

-0.06

Talking

-0.06

;">

-0.06

ledon

-0.06

оба

-0.06

 isolate

-0.06

 رود

-0.06

POSITIVE LOGITS

"http

0.07

 قاب

0.07

 Construct

0.07

	Context

0.06

 Relax

0.06

 discern

0.06

 عفش

0.06

 Cement

0.06

 asses

0.06

ชม

0.06

Activations Density 0.005%

characters

The neuron activates on meta‐narrative tokens—words that refer to the fictional framing or characters (e.g. “character,” “characters,” “fiction,” “scenario”).

No Comments

No Known Activations