INDEX

Explanations

terms related to mathematical operators and matrices

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

irit

-0.14

emoc

-0.14

è¡Ľ

-0.14

 hete

-0.14

 wart

-0.14

 Rout

-0.14

taj

-0.14

odom

-0.14

lags

-0.13

sex

-0.13

POSITIVE LOGITS

 dictionary

0.30

 recovery

0.30

 Dictionary

0.28

 dictionaries

0.27

 reconstruction

0.27

Dictionary

0.27

dictionary

0.27

 Recovery

0.27

 sparse

0.26

 compress

0.24

Activations Density 0.008%

terms related to mathematical operators and matrices

No Comments

No Known Activations