INDEX

Explanations

references to individual names and titles

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.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

 Karel

-0.18

 Voor

-0.17

erman

-0.16

 Powers

-0.16

Gew

-0.15

 Morse

-0.15

çĳŀ

-0.15

utch

-0.15

ppy

-0.14

azor

-0.14

POSITIVE LOGITS

iÃŁ

0.20

 Sting

0.19

-Christian

0.19

 Broker

0.19

 Diet

0.18

 Dipl

0.17

 Becker

0.17

asers

0.16

 Loch

0.16

 Bless

0.16

Activations Density 0.141%

references to individual names and titles

No Comments

No Known Activations