INDEX

Explanations

intensifiers that enhance the attributes of nouns or actions

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

Not in Any Lists

No Comments

Negative Logits

 better

-0.16

irk

-0.15

 Better

-0.15

sic

-0.15

kip

-0.15

osas

-0.15

ict

-0.14

 assez

-0.14

 more

-0.14

 meilleur

-0.14

POSITIVE LOGITS

much

0.31

 much

0.31

 Much

0.25

Much

0.25

 MUCH

0.21

yyyy

0.20

yy

0.18

yyy

0.18

DOC

0.18

 seldom

0.17

Activations Density 0.084%

intensifiers that enhance the attributes of nouns or actions

No Comments

No Known Activations