INDEX

Explanations

Internet forum comments

np_max-act · gemini-2.0-flash

This neuron responds to the model’s special control/formatting tokens (e.g. start/end markers like `<s>`, `</s>`, `[INST]`, header IDs), not to ordinary words in the content.

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

‌ن

-0.06

amb

-0.06

alm

-0.06

CanBeConverted

-0.06

realm

-0.06

_mD

-0.06

itors

-0.06

ετ

-0.06

panel

-0.06

daughter

-0.06

POSITIVE LOGITS

}");↵

0.07

Cs

0.06

�

0.06

_sg

0.06

radan

0.06

 Со

0.06

_definitions

0.06

.$

0.06

<Input

0.06

Setter

0.06

Activations Density 0.005%

Internet forum comments

This neuron responds to the model’s special control/formatting tokens (e.g. start/end markers like `<s>`, `</s>`, `[INST]`, header IDs), not to ordinary words in the content.

No Comments

No Known Activations