INDEX

Explanations

phrases related to conflict or critique

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

okud

-0.17

adol

-0.17

ickers

-0.16

bes

-0.15

emd

-0.15

tru

-0.15

ustil

-0.14

nin

-0.14

åªĴ

-0.14

iba

-0.14

POSITIVE LOGITS

 upon

0.59

on

0.53

 Upon

0.39

äºİ

0.37

upon

0.36

 Ø¹ÙĦÙī

0.35

Upon

0.34

 Ð½Ð°

0.34

 pada

0.33

 pÃ¥

0.32

Activations Density 0.184%

phrases related to conflict or critique

No Comments

No Known Activations