INDEX

Explanations

references to specific geographic locations and capitals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 Croatia

-0.16

utz

-0.16

ocal

-0.15

 Yugoslavia

-0.15

 Neville

-0.14

 Catalonia

-0.14

orthy

-0.14

 Louisiana

-0.14

Descriptor

-0.14

 Ventura

-0.14

POSITIVE LOGITS

 Islamabad

0.28

 capitals

0.23

 Ankara

0.23

 Kabul

0.22

 capital

0.22

 Moscow

0.20

 Pret

0.20

 Capitals

0.20

 Jakarta

0.19

 Canberra

0.19

Activations Density 0.185%

references to specific geographic locations and capitals

No Comments

No Known Activations