INDEX

Explanations

names and titles associated with individuals in a corporate or organizational context

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 21-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L21R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.21.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

 Latina

-0.15

_mgr

-0.15

lord

-0.15

ehir

-0.14

 Dame

-0.13

 refs

-0.13

andal

-0.13

Ã§e

-0.13

shima

-0.13

 geometry

-0.13

POSITIVE LOGITS

 former

0.19

 author

0.17

 Former

0.16

alias

0.16

Former

0.16

 Owner

0.15

.GetObject

0.15

 Founder

0.15

 member

0.15

esiz

0.15

Activations Density 0.081%

names and titles associated with individuals in a corporate or organizational context

No Comments

No Known Activations