INDEX

Explanations

references to royal courts or similar hierarchical structures

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

ovit

-0.19

hausen

-0.18

erator

-0.16

icie

-0.16

rant

-0.15

dorf

-0.15

Ð´Ðµ

-0.15

éĥİ

-0.14

aukee

-0.14

hart

-0.14

POSITIVE LOGITS

neys

0.18

sey

0.16

aller

0.14

sov

0.14

UBLE

0.14

lys

0.13

ly

0.13

angel

0.13

reesome

0.13

 Cust

0.13

Activations Density 0.028%

references to royal courts or similar hierarchical structures

No Comments

No Known Activations