INDEX

Explanations

references to time-related phrases

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 24-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L24R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.24.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

ettel

-0.19

redit

-0.15

ivil

-0.15

Ã¹ng

-0.15

uttle

-0.15

olly

-0.15

ducted

-0.14

å¾Ĵ

-0.14

ewith

-0.14

jekt

-0.14

POSITIVE LOGITS

here

0.23

ruz

0.23

inski

0.22

lan

0.19

 Trend

0.19

 Playing

0.18

 Hiring

0.18

withstanding

0.17

ayout

0.17

HERE

0.17

Activations Density 0.035%

references to time-related phrases

No Comments

No Known Activations