INDEX

Explanations

references to ports and portraits

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

-quarters

-0.19

elda

-0.19

itch

-0.16

plnÄĽ

-0.16

asing

-0.16

sei

-0.16

ÛĮØ´

-0.16

ä¿Ĺ

-0.15

elder

-0.15

éĥİ

-0.15

POSITIVE LOGITS

folios

0.27

raits

0.26

ugal

0.26

smouth

0.24

side

0.21

æ¹¾

0.21

ability

0.21

age

0.20

çģ£

0.17

une

0.17

Activations Density 0.066%

references to ports and portraits

No Known Activations