INDEX

Explanations

references to academic conferences and publications

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

ensis

-0.16

ÑīÐµ

-0.15

[".

-0.15

Ø´Øª

-0.14

hog

-0.14

elage

-0.14

ÑģÑĤÐ²Ð¾ÑĢ

-0.14

ukt

-0.14

Ã¼sÃ¼

-0.14

ledon

-0.13

POSITIVE LOGITS

ACM

0.26

AAA

0.20

IEEE

0.19

 IEEE

0.19

 workshop

0.19

.sig

0.19

acl

0.18

sig

0.18

Dag

0.18

AC

0.18

Activations Density 0.058%

references to academic conferences and publications

No Comments

No Known Activations