INDEX

Explanations

the word "or" used to present alternatives or choices

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

educt

-0.18

ubern

-0.17

udeau

-0.17

ilation

-0.17

uzey

-0.16

/or

-0.16

eraÃ§Ã£o

-0.16

orgia

-0.15

leo

-0.15

 Briggs

-0.15

POSITIVE LOGITS

 else

0.20

phans

0.17

anges

0.17

chest

0.17

 rather

0.16

 indeed

0.16

naments

0.16

ner

0.15

ote

0.15

 simply

0.15

Activations Density 0.071%

the word "or" used to present alternatives or choices

No Comments

No Known Activations