INDEX

Explanations

phrases that express a sense of moderation or condition emphasizing "pretty much."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

ulumi

-0.17

hma

-0.15

mars

-0.15

ither

-0.15

sts

-0.14

 Very

-0.14

omer

-0.14

atee

-0.14

odelist

-0.14

egis

-0.14

POSITIVE LOGITS

 much

0.41

much

0.30

 Much

0.29

Much

0.27

 darn

0.26

 MUCH

0.26

 damn

0.21

-ÑĤÐ°ÐºÐ¸

0.21

ing

0.21

-boy

0.20

Activations Density 0.018%

phrases that express a sense of moderation or condition emphasizing "pretty much."

No Comments

No Known Activations

phrases that express a sense of moderation or condition emphasizing "pretty much."

No Comments

No Known Activations