INDEX

Explanations

references to film titles and awards

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

 Fashion

-0.15

ÑĽ

-0.15

esser

-0.15

 Leban

-0.14

ISBN

-0.14

porter

-0.14

tura

-0.14

arem

-0.13

iar

-0.13

ÅĻ

-0.13

POSITIVE LOGITS

foi

0.29

te

0.26

era

0.23

 estava

0.23

 tinha

0.21

itou

0.19

 recebe

0.19

fe

0.19

 trou

0.19

fic

0.19

Activations Density 0.008%

references to film titles and awards

No Known Activations