INDEX

Explanations

terms related to levels of quality or measurement

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

 enough

-0.27

 Enough

-0.22

ish

-0.19

umi

-0.19

er

-0.17

ers

-0.16

isher

-0.15

Enough

-0.15

Ø§

-0.15

ij

-0.15

POSITIVE LOGITS

-ever

0.44

 possible

0.38

possible

0.35

 Possible

0.31

ablish

0.30

Possible

0.29

 posible

0.29

 imaginable

0.28

 possibile

0.28

 moÅ¾nÃ©

0.26

Activations Density 0.067%

terms related to levels of quality or measurement

No Comments

No Known Activations