INDEX

Explanations

questions and punctuation marks, particularly in conversational contexts

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

ipur

-0.17

arÃ¡

-0.15

ujte

-0.14

 Ð¾Ð¶

-0.14

isode

-0.14

 Worth

-0.14

ãģ«ãģ¤

-0.14

_expected

-0.14

Clr

-0.14

putc

-0.13

POSITIVE LOGITS

 Want

0.39

Want

0.38

 want

0.36

 Need

0.36

Are

0.35

 need

0.34

Need

0.34

want

0.33

Did

0.32

Are

0.31

Activations Density 0.173%

questions and punctuation marks, particularly in conversational contexts

No Comments

No Known Activations