INDEX

Explanations

words related to arguments and debates

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

eh

-0.16

closure

-0.15

chen

-0.15

 closure

-0.14

Ult

-0.14

Î¶Î±

-0.14

igkeit

-0.14

ral

-0.14

 Sachs

-0.14

ycz

-0.14

POSITIVE LOGITS

uably

0.39

entin

0.35

entine

0.33

onaut

0.31

uing

0.30

inine

0.29

yle

0.29

ÑĥÐ¼ÐµÐ½ÑĤ

0.28

entina

0.28

uable

0.28

Activations Density 0.007%

words related to arguments and debates

No Comments

No Known Activations