INDEX

Explanations

phrases related to being overlooked or unnoticed

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

_UNUSED

-0.16

ãĤį

-0.15

Ð¾ÑĹ

-0.15

ahas

-0.15

ÏĥÏĢ

-0.14

ifacts

-0.14

omial

-0.14

gew

-0.14

íĥ

-0.13

/package

-0.13

POSITIVE LOGITS

 radar

0.45

 Radar

0.33

rad

0.28

 flying

0.25

fly

0.25

ä½İ

0.23

fly

0.21

adar

0.21

RAD

0.21

Fly

0.21

Activations Density 0.069%

phrases related to being overlooked or unnoticed

No Comments

No Known Activations