INDEX

Explanations

expressions related to humorous or playful scenarios

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

 while

-0.15

rine

-0.15

 since

-0.14

ÑĩÐµ

-0.14

Additionally

-0.13

oggle

-0.13

Gee

-0.13

 however

-0.13

 although

-0.13

 Additionally

-0.13

POSITIVE LOGITS

And

0.38

And

0.31

Or

0.23

 Including

0.21

_and

0.20

 Which

0.20

and

0.20

 åĴĮ

0.19

Or

0.19

AND

0.19

Activations Density 0.307%

expressions related to humorous or playful scenarios

No Comments

No Known Activations