INDEX

Explanations

expressions that convey a degree of emphasis or moderation

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 31-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L31R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.31.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

 very

-0.20

 Very

-0.19

 exceedingly

-0.18

 ráº¥t

-0.18

 extremely

-0.17

 VERY

-0.17

 Extremely

-0.17

very

-0.17

 really

-0.17

 somewhat

-0.17

POSITIVE LOGITS

 much

0.35

 darn

0.31

 Much

0.26

much

0.24

 dang

0.24

 damn

0.24

Much

0.24

-boy

0.23

-m

0.23

 damned

0.21

Activations Density 0.025%

expressions that convey a degree of emphasis or moderation

No Comments

No Known Activations