INDEX

Explanations

intensifiers or modifiers that emphasize the degree of something, particularly the word "very"

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

irk

-0.17

olt

-0.16

ital

-0.15

ense

-0.14

ery

-0.14

ict

-0.14

igor

-0.14

aroo

-0.13

IQ

-0.13

vig

-0.13

POSITIVE LOGITS

yyyy

0.20

yyy

0.17

ocoder

0.17

yy

0.16

/ext

0.15

 much

0.15

eur

0.15

uger

0.14

angel

0.14

yl

0.14

Activations Density 0.062%

intensifiers or modifiers that emphasize the degree of something, particularly the word "very"

No Comments

No Known Activations