INDEX

Explanations

various forms of the word "something" indicating the discussion of undefined or unspecified concepts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 25-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L25R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.25.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

 somehow

-0.17

ØŃÙĬ

-0.15

 anything

-0.14

sync

-0.14

à¸ŀà¸Ń

-0.14

abeth

-0.13

ikel

-0.13

ÚĨÙĩ

-0.13

éĤ£ç§į

-0.13

same

-0.13

POSITIVE LOGITS

 else

0.45

else

0.32

_else

0.29

Else

0.29

 ELSE

0.26

	else

0.26

 Else

0.25

ELSE

0.21

/s

0.20

ness

0.19

Activations Density 0.054%

various forms of the word "something" indicating the discussion of undefined or unspecified concepts

No Comments

No Known Activations