INDEX

Explanations

concepts related to simplicity or ease

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 2-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L2R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.2.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

 Bain

-0.16

 gain

-0.16

 base

-0.15

joy

-0.15

-0.14

 BASE

-0.14

ãĥĥ

-0.14

hic

-0.14

ascus

-0.14

 model

-0.14

POSITIVE LOGITS

asal

0.19

asper

0.15

ight

0.15

oro

0.14

marks

0.14

Mob

0.14

anger

0.14

falls

0.14

maj

0.14

mlink

0.14

Activations Density 0.060%

concepts related to simplicity or ease

No Comments

No Known Activations