INDEX

Explanations

references to quantities or sequences involving the number six

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

ont

-0.17

aren

-0.17

ived

-0.16

iams

-0.15

lest

-0.15

ishments

-0.15

gram

-0.15

led

-0.15

action

-0.14

POSITIVE LOGITS

teenth

0.37

teen

0.32

ties

0.31

ti

0.28

ty

0.27

sense

0.21

TY

0.20

tee

0.19

tel

0.18

iang

0.18

Activations Density 0.111%

references to quantities or sequences involving the number six

No Comments

No Known Activations