INDEX

Explanations

punctuation and expressions of strong opinions

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 14-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L14R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.14.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

veloper

-0.17

 Conv

-0.15

Gate

-0.14

FileSystem

-0.14

 (*((

-0.14

Ark

-0.14

ilon

-0.14

èµ·

-0.14

edo

-0.14

 åĲī

-0.13

POSITIVE LOGITS

 seal

0.15

opot

0.15

 Seal

0.15

ukan

0.14

urus

0.14

ickle

0.14

ksen

0.14

Ã¼nd

0.14

 factorial

0.14

 Bundes

0.14

Activations Density 0.001%

punctuation and expressions of strong opinions

No Comments

No Known Activations