INDEX

Explanations

instances of the word "who."

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 29-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L29R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.29.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

robat

-0.20

ted

-0.17

 Darling

-0.17

bian

-0.16

mented

-0.15

bol

-0.15

cline

-0.15

aises

-0.15

nt

-0.15

net

-0.14

POSITIVE LOGITS

ops

0.30

ever

0.30

 else

0.28

opi

0.26

oping

0.26

osh

0.25

am

0.23

op

0.22

opsy

0.22

ope

0.21

Activations Density 0.025%

instances of the word "who."

No Comments

No Known Activations