INDEX

Explanations

references to car features and design elements

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

Ð±Ð¾ÑĤ

-0.17

Outer

-0.16

 Outer

-0.15

 Attach

-0.15

wright

-0.14

æĤł

-0.14

idis

-0.14

 Chamber

-0.14

ernel

-0.14

 attachment

-0.14

POSITIVE LOGITS

 dash

0.28

 dashboard

0.26

 Dashboard

0.23

 Dash

0.21

dashboard

0.21

 cockpit

0.20

Dash

0.20

Dashboard

0.20

.dashboard

0.19

 steering

0.19

Activations Density 0.062%

references to car features and design elements

No Comments

No Known Activations