INDEX

Explanations

phrases related to training and education

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

_training

-0.32

 trained

-0.31

 training

-0.31

_train

-0.31

 Training

-0.28

_TRAIN

-0.28

Training

-0.28

training

-0.25

trained

-0.23

è®Ń

-0.23

POSITIVE LOGITS

ees

0.30

ee

0.30

 wheels

0.21

 Wheels

0.20

ning

0.20

ings

0.19

ining

0.18

/testing

0.17

 session

0.16

 sessions

0.16

Activations Density 0.037%

phrases related to training and education

No Comments

No Known Activations