INDEX

Explanations

references to web-related terms and entities

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

Not in Any Lists

No Comments

Negative Logits

à¸£à¸Ķ

-0.18

edges

-0.18

WebService

-0.17

leo

-0.17

ances

-0.17

overe

-0.16

fy

-0.16

ses

-0.15

WebKit

-0.15

arium

-0.15

POSITIVE LOGITS

inars

0.33

inar

0.28

cams

0.26

iste

0.25

ber

0.24

isodes

0.24

nesday

0.24

isode

0.24

bing

0.24

.archive

0.23

Activations Density 0.027%

references to web-related terms and entities

No Comments

No Known Activations