INDEX

Explanations

sequences of numbers, particularly related to historical or academic references

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

iron

-0.15

 direct

-0.15

erap

-0.15

utdown

-0.15

UiThread

-0.14

irect

-0.14

pom

-0.14

ãĥ³ãĥ

-0.14

hend

-0.14

POSITIVE LOGITS

å·±

0.16

Ã©ri

0.15

ff

0.15

æ¿

0.14

èİİ

0.14

 ÄĳÃ¢u

0.14

.minute

0.14

oves

0.14

idor

0.14

ANJI

0.14

Activations Density 0.030%

sequences of numbers, particularly related to historical or academic references

No Comments

No Known Activations