INDEX

Explanations

references to academic conferences and their associated proceedings

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

hower

-0.17

 earliest

-0.15

Ã²n

-0.14

 Interview

-0.14

hem

-0.14

 Slate

-0.14

 reform

-0.13

ivated

-0.13

ered

-0.13

erd

-0.13

POSITIVE LOGITS

 proceedings

0.33

Proceed

0.28

 Proceedings

0.28

 Proceed

0.28

 proceeding

0.27

 Workshop

0.23

 workshop

0.22

 proceed

0.22

 proced

0.22

 proceeds

0.21

Activations Density 0.083%

references to academic conferences and their associated proceedings

No Known Activations