INDEX

Explanations

sequences of code or mathematical notations

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

Not in Any Lists

No Comments

Negative Logits

-0.86

Ùĩ

-0.45

ska

-0.39

sak

-0.37

sian

-0.32

sage

-0.31

sip

-0.31

à¸Ĺ

-0.30

POSITIVE LOGITS

raquo

0.17

emsp

0.16

eyh

0.16

 saja

0.16

Ð¾ÑĢÑĸ

0.15

'gc

0.14

.Ultra

0.14

ãĥ³ãĥĦ

0.14

ratulations

0.14

-unused

0.13

Activations Density 0.142%

sequences of code or mathematical notations

No Comments

No Known Activations