INDEX

Explanations

Quotation marks

np_max-act · gemini-2.0-flash

This neuron detects configuration or code constant tokens—uppercase or enum‐style identifiers (e.g. enum names or option values) in code/config snippets.

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

رد

-0.07

λίου

-0.07

neighbor

-0.07

เกล

-0.07

usalem

-0.06

ites

-0.06

igt

-0.06

amic

-0.06

 Indonesia

-0.06

 chậm

-0.06

POSITIVE LOGITS

XB

0.07

Reflect

0.07

STA

0.06

$arr

0.06

 Reflect

0.06

誠

0.06

/int

0.06

 กำ

0.06

rk

0.06

ีย

0.06

Activations Density 0.052%

Quotation marks

This neuron detects configuration or code constant tokens—uppercase or enum‐style identifiers (e.g. enum names or option values) in code/config snippets.

No Comments

No Known Activations