INDEX

Explanations

characters and symbols related to specific content formats or encoding issues

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

tero

-0.14

ufen

-0.14

izo

-0.14

ittel

-0.14

omo

-0.14

ysz

-0.14

iran

-0.13

OrFail

-0.13

'gc

-0.13

POSITIVE LOGITS

 account

0.20

 Account

0.19

account

0.19

 compte

0.18

Account

0.18

mil

0.16

 Accounts

0.16

.Account

0.16

 accounts

0.16

_account

0.15

Activations Density 0.005%

characters and symbols related to specific content formats or encoding issues

No Comments

No Known Activations