INDEX

Explanations

specific character sequences or patterns within words, likely focusing on names or titles

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

ted

-0.17

de

-0.16

Ã¤n

-0.15

 NÃ¤

-0.14

Ã¤s

-0.14

Ã¤re

-0.14

ler

-0.14

um

-0.14

 nich

-0.14

ÃŃa

-0.14

POSITIVE LOGITS

.hu

0.21

cs

0.21

 Hungary

0.19

-Cs

0.18

csi

0.18

 Hungarian

0.18

elman

0.18

Hung

0.18

Sz

0.18

º

0.17

Activations Density 0.002%

specific character sequences or patterns within words, likely focusing on names or titles

No Comments

No Known Activations