INDEX

Explanations

references to historical artifacts and their descriptions

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

 gravel

-0.17

eva

-0.16

 monument

-0.15

arse

-0.15

Ãłu

-0.15

dum

-0.15

ÑģÑĤÐ¸Ð½

-0.14

esar

-0.14

inkel

-0.14

elves

-0.13

POSITIVE LOGITS

 maps

0.26

 Maps

0.23

 collect

0.22

maps

0.22

 memor

0.20

 rare

0.20

 framed

0.20

 ephem

0.19

aut

0.19

 scrim

0.19

Activations Density 0.221%

references to historical artifacts and their descriptions

No Comments

No Known Activations