INDEX

Explanations

phrases and terms that indicate a continuation or follow-up in discussions or narratives

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

okers

-0.16

uts

-0.15

atis

-0.15

Î»Ïħ

-0.15

ä¸Ģé¡µ

-0.14

å¡ļ

-0.14

URED

-0.13

dling

-0.13

aho

-0.13

Knife

-0.13

POSITIVE LOGITS

-up

0.42

up

0.34

-Up

0.32

up

0.31

-ups

0.25

UP

0.23

(up

0.21

Up

0.21

.up

0.20

_up

0.20

Activations Density 0.010%

phrases and terms that indicate a continuation or follow-up in discussions or narratives

No Comments

No Known Activations