INDEX

Explanations

many

np_max-act · gemini-2.0-flash

The neuron activates on placeholder tokens representing character names (e.g. “NAME_1,” “NAME_2,” etc.) in the text.

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

 اصلی

-0.06

 праців

-0.06

 Wood

-0.06

rb

-0.06

 milieu

-0.06

-controlled

-0.06

 Steph

-0.06

Wrap

-0.06

edo

-0.06

oruč

-0.06

POSITIVE LOGITS

_dirty

0.07

LIK

0.07

-grey

0.06

.Groups

0.06

 BASIC

0.06

 antim

0.06

 Repos

0.06

 پاس

0.06

 volley

0.06

province

0.06

Activations Density 0.111%

many

The neuron activates on placeholder tokens representing character names (e.g. “NAME_1,” “NAME_2,” etc.) in the text.

No Comments

No Known Activations