INDEX

Explanations

references to NBC and its programs or hosts

oai_token-act-pair · gpt-4o-mini

Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.hook_resid_post

Hook Layer

Architecture

jumprelu

Context Size

1,024

Dataset

cerebras/SlimPajama-627B

Activation Function

relu

Embeds

IFrame

Link

Negative Logits

åĹ

-0.17

 Fairfax

-0.16

_IW

-0.15

_PIX

-0.15

armacy

-0.14

.Std

-0.14

ÃŃv

-0.14

ddb

-0.14

æ¡Ĳ

-0.14

utura

-0.14

POSITIVE LOGITS

NBC

0.36

NBC

0.35

Universal

0.29

 Universal

0.25

 TODAY

0.25

 UNIVERS

0.24

 universal

0.23

universal

0.23

NB

0.21

nb

0.19

Activations Density 0.036%

references to NBC and its programs or hosts

No Known Activations