INDEX

Explanations

mathematical terms and references, particularly related to equations and physical principles

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

nees

-0.15

 Victim

-0.14

 gauss

-0.14

 Laud

-0.14

emean

-0.14

zik

-0.13

 Voter

-0.13

_portal

-0.13

mor

-0.13

 Veteran

-0.13

POSITIVE LOGITS

va

0.55

vo

0.51

ve

0.50

vi

0.49

view

0.41

ves

0.39

vem

0.39

ven

0.39

0.36

Activations Density 0.100%

mathematical terms and references, particularly related to equations and physical principles

No Comments

No Known Activations