INDEX

Explanations

calls to action for clicking links

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

aniel

-0.17

/by

-0.16

stav

-0.16

ynes

-0.16

pag

-0.15

ptions

-0.14

 Wake

-0.14

dal

-0.14

icipants

-0.14

uling

-0.14

POSITIVE LOGITS

ety

0.27

tiv

0.20

through

0.20

hole

0.20

.echo

0.19

/t

0.19

-through

0.19

inks

0.17

 HERE

0.17

ETY

0.17

Activations Density 0.024%

calls to action for clicking links

No Comments

No Known Activations