INDEX

Explanations

select

np_max-act · gemini-2.0-flash

The neuron activates on SQL code tokens—especially keywords and constructs in example queries (e.g. SELECT, OVER, PARTITION BY, ROW_NUMBER).

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

	when

-0.06

 mediator

-0.06

로서

-0.06

 where

-0.06

Means

-0.06

 Hairst

-0.06

.books

-0.05

xs

-0.05

Wars

-0.05

แล

-0.05

POSITIVE LOGITS

│

0.07

_DECL

0.07

(UnmanagedType

0.07

 edil

0.06

�

0.06

 перев

0.06

oenix

0.06

/>\

0.06

Discovery

0.06

 TForm

0.06

Activations Density 0.003%

select

The neuron activates on SQL code tokens—especially keywords and constructs in example queries (e.g. SELECT, OVER, PARTITION BY, ROW_NUMBER).

No Comments

No Known Activations