INDEX

Explanations

references to keys and key-related terminology

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

jc

-0.17

ationToken

-0.16

al

-0.16

é»ĺ

-0.15

pus

-0.15

ickers

-0.15

apia

-0.15

sky

-0.15

 heck

-0.15

alley

-0.15

POSITIVE LOGITS

hole

0.35

chain

0.33

chains

0.27

ring

0.26

holes

0.24

logger

0.24

cloak

0.23

note

0.23

tar

0.23

boards

0.23

Activations Density 0.056%

references to keys and key-related terminology

No Comments

No Known Activations