INDEX

Explanations

contractions involving "are" and "you're."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 27-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L27R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.27.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

xCD

-0.16

ilon

-0.16

 itself

-0.15

 çĿ

-0.14

idir

-0.14

sume

-0.14

PEND

-0.13

rw

-0.13

ivan

-0.13

POSITIVE LOGITS

 lucky

0.23

 ever

0.22

nt

0.20

 fortunate

0.18

 unfamiliar

0.18

 anywhere

0.17

 unsure

0.17

 planning

0.17

 feeling

0.17

ucky

0.16

Activations Density 0.075%

contractions involving "are" and "you're."

No Comments

No Known Activations