INDEX

Explanations

.

np_max-act · gemini-2.0-flash

This neuron activates on boilerplate patent-style transition phrases (e.g. “Furthermore,” “will,” “become,” “apparent”) that introduce the detailed description or appended claims.

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

lád

-0.07

tf

-0.07

gov

-0.06

 hero

-0.06

‌دهد

-0.06

 Objective

-0.06

ECS

-0.06

 skipping

-0.06

 radiation

-0.06

nist

-0.06

POSITIVE LOGITS

//}}

0.07

 ном

0.06

	mkdir

0.06

 Spear

0.06

 лич

0.06

">'+

0.06

Plug

0.06

.boolean

0.06

ley

0.06

科技

0.06

Activations Density 0.006%

.

This neuron activates on boilerplate patent-style transition phrases (e.g. “Furthermore,” “will,” “become,” “apparent”) that introduce the detailed description or appended claims.

No Comments

No Known Activations