INDEX

Explanations

terms related to mathematical representations and structures

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

 Bain

-0.18

Ace

-0.17

 trot

-0.17

fall

-0.17

ohl

-0.16

ker

-0.16

APPER

-0.16

 Signature

-0.16

ace

-0.15

uby

-0.15

POSITIVE LOGITS

Aus

0.20

regor

0.20

Aus

0.19

 Happ

0.17

adero

0.17

 Heller

0.17

 Blocks

0.16

Naw

0.16

 simples

0.16

 blocks

0.16

Activations Density 0.026%

terms related to mathematical representations and structures

No Comments

No Known Activations