INDEX

Explanations

terms related to software errors and trust relationships in user accounts

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

 Huffman

-0.16

Lon

-0.15

urat

-0.15

 Hers

-0.14

SDL

-0.14

 Cust

-0.14

 Burr

-0.14

ãĥ³ãĥĶ

-0.14

 cust

-0.14

piring

-0.13

POSITIVE LOGITS

ker

0.34

Ker

0.31

 ticket

0.30

 Ticket

0.28

 tickets

0.27

Ticket

0.27

-ticket

0.27

ticket

0.27

ker

0.25

 Princip

0.25

Activations Density 0.013%

terms related to software errors and trust relationships in user accounts

No Comments

No Known Activations