INDEX

Explanations

don

np_max-act · gemini-2.0-flash

This neuron detects the special marker tokens used for headers and metadata (e.g. <|start_header_id|> and <|end_header_id|>) in the conversation transcript.

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

MainThread

-0.07

식

-0.07

没

-0.07

 القرن

-0.06

 bulunmaktadır

-0.06

 noon

-0.06

tons

-0.06

ówn

-0.06

\R

-0.06

※

-0.06

POSITIVE LOGITS

 compliment

0.07

 Blocked

0.06

 bloque

0.06

arrant

0.06

vyšší

0.06

 уровне

0.06

memset

0.06

	PreparedStatement

0.06

 queried

0.06

िकट

0.06

Activations Density 0.046%

don

This neuron detects the special marker tokens used for headers and metadata (e.g. <|start_header_id|> and <|end_header_id|>) in the conversation transcript.

No Comments

No Known Activations