INDEX

Explanations

terms related to euphemistic language and societal attitudes towards sensitive topics

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 9-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L9R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.9.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

ernal

-0.16

agrid

-0.16

erior

-0.16

abit

-0.15

Ã¸re

-0.15

èħ¹

-0.14

chrift

-0.14

åŁºåľ°

-0.13

Âİ

-0.13

 Tradable

-0.13

POSITIVE LOGITS

 terms

0.18

terms

0.16

 Terms

0.16

æľ¯

0.16

ucci

0.16

åĭ

0.16

umbo

0.15

 jenter

0.15

Term

0.15

Rut

0.15

Activations Density 0.334%

terms related to euphemistic language and societal attitudes towards sensitive topics

No Comments

No Known Activations