INDEX

Explanations

programming code structures and keywords

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 18-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L18R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.18.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

 Bour

-0.17

onom

-0.16

tsy

-0.15

 theano

-0.15

jure

-0.14

ìĦ¸

-0.14

 Truy

-0.14

 creep

-0.14

 Benghazi

-0.14

onso

-0.14

POSITIVE LOGITS

 Bart

0.18

 bart

0.17

bart

0.17

idden

0.16

MLM

0.15

 Lightning

0.15

 lightning

0.15

tae

0.14

Sau

0.14

oyer

0.14

Activations Density 0.000%

programming code structures and keywords

No Comments

No Known Activations