INDEX

Explanations

variations of the prefix "ped," suggesting a focus on terms related to children or education

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

edImage

-0.18

yz

-0.17

ynes

-0.16

TEL

-0.15

iners

-0.15

 amazon

-0.14

ARIO

-0.14

merge

-0.14

IEW

-0.14

xCD

-0.14

POSITIVE LOGITS

agog

0.32

antic

0.30

ersen

0.28

iatric

0.28

iatrics

0.27

dling

0.25

antry

0.24

est

0.24

dle

0.23

dl

0.23

Activations Density 0.007%

variations of the prefix "ped," suggesting a focus on terms related to children or education

No Comments

No Known Activations