INDEX

Explanations

words and phrases related to manipulation and control

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

ungan

-0.15

lice

-0.14

Ø®Øµ

-0.14

 enrich

-0.14

eno

-0.14

 Roger

-0.14

rei

-0.14

Me

-0.14

OfWork

-0.14

POSITIVE LOGITS

wald

0.18

å¤ªéĥİ

0.17

ÙĴØ¹

0.16

relude

0.15

eo

0.15

åĢį

0.15

egend

0.15

 Horton

0.15

æļ®

0.15

Activations Density 0.007%

words and phrases related to manipulation and control

No Comments

No Known Activations