INDEX

Explanations

punctuation

np_max-act · gemini-2.0-flash

This neuron detects judge attribution lines—i.e. the names and initials of judges (e.g. “Smith, J.” or “Jones, JJ.”) at the ends of opinions.

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

 podmínky

-0.07

cmpeq

-0.07

 squeez

-0.07

ันย

-0.07

与

-0.06

 Uganda

-0.06

 môi

-0.06

 nostalgic

-0.06

 баз

-0.06

 thử

-0.06

POSITIVE LOGITS

diag

0.08

№№

0.07

locate

0.07

 glam

0.06

 hồ

0.06

�

0.06

 ascertain

0.06

 sofa

0.06

Gro

0.06

-bot

0.06

Activations Density 0.006%

punctuation

This neuron detects judge attribution lines—i.e. the names and initials of judges (e.g. “Smith, J.” or “Jones, JJ.”) at the ends of opinions.

No Comments

No Known Activations

punctuation

This neuron detects judge attribution lines—i.e. the names and initials of judges (e.g. “Smith, J.” or “Jones, JJ.”) at the ends of opinions.

No Comments

No Known Activations