INDEX

Explanations

Numbers and code

np_max-act · gemini-2.0-flash

The neuron fires on numeric tokens—especially counts or limits—whenever the user specifies a quantity (e.g. “15” in “less than 15 characters”).

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

øre

-0.07

.bar

-0.07

Kor

-0.06

 svém

-0.06

ěn

-0.06

Inserted

-0.06

 признач

-0.06

 παιδ

-0.06

 homogeneous

-0.06

.utils

-0.06

POSITIVE LOGITS

 Nicole

0.07

IPH

0.07

>",↵

0.06

 HIGH

0.06

�

0.06

:NS

0.06

 ||↵

0.06

_THRESH

0.06

levels

0.06

<Node

0.06

Activations Density 0.062%

Numbers and code

The neuron fires on numeric tokens—especially counts or limits—whenever the user specifies a quantity (e.g. “15” in “less than 15 characters”).

No Comments

No Known Activations

Numbers and code

The neuron fires on numeric tokens—especially counts or limits—whenever the user specifies a quantity (e.g. “15” in “less than 15 characters”).

No Comments

No Known Activations