INDEX

Explanations

data

np_max-act · gemini-2.0-flash

statements and concepts related to few-shot learning techniques, particularly in the context of machine learning models and their efficiency.

oai_token-act-pair · gpt-4o-mini Triggered by @xinyanhu8

This neuron detects mentions of few-shot learning, especially references to training with a very small number of labeled examples.

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

 systemFontOfSize

-0.07

 tattoo

-0.06

 Ayrıca

-0.06

 Russia

-0.06

ことを

-0.06

-exec

-0.06

杜

-0.06

новаж

-0.06

-bin

-0.06

Thomas

-0.06

POSITIVE LOGITS

 Portsmouth

0.07

ổi

0.07

Sql

0.07

 ReturnValue

0.07

 DefaultValue

0.07

енные

0.07

 Residential

0.07

 consideration

0.06

_horizontal

0.06

Activations Density 0.016%

data

statements and concepts related to few-shot learning techniques, particularly in the context of machine learning models and their efficiency.

This neuron detects mentions of few-shot learning, especially references to training with a very small number of labeled examples.

No Comments

No Known Activations