INDEX

Explanations

phrases that involve associations or relationships

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

uos

-0.15

Nob

-0.15

dob

-0.15

èı

-0.14

RAND

-0.14

uv

-0.13

prit

-0.13

ophe

-0.13

onn

-0.13

cci

-0.13

POSITIVE LOGITS

 regards

0.21

 regard

0.19

 respecto

0.19

tÃ¼r

0.15

ReadWrite

0.15

uten

0.15

loh

0.15

artial

0.14

supported

0.14

olle

0.14

Activations Density 0.155%

phrases that involve associations or relationships

No Comments

No Known Activations