INDEX

Explanations

comparative terms related to improvement or quality

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

arna

-0.15

elige

-0.14

Ð´Ð°Ð»

-0.14

adem

-0.14

uff

-0.14

Ã§i

-0.14

AndPassword

-0.14

ungan

-0.13

ieder

-0.13

POSITIVE LOGITS

-than

0.39

ment

0.35

 than

0.35

idge

0.29

-known

0.29

than

0.29

_than

0.29

ing

0.27

Than

0.27

 Than

0.26

Activations Density 0.039%

comparative terms related to improvement or quality

No Comments

No Known Activations