INDEX

Explanations

references to repeated actions or processes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 30-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L30R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.30.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

 Restart

-0.17

Restart

-0.17

latest

-0.16

amburger

-0.16

finally

-0.16

/latest

-0.16

onse

-0.15

ardash

-0.15

once

-0.14

Ascii

-0.14

POSITIVE LOGITS

ag

0.57

 gain

0.54

gain

0.48

Gain

0.44

 Gain

0.44

.ag

0.42

 again

0.42

_gain

0.39

ag

0.37

 gains

0.36

Activations Density 0.049%

references to repeated actions or processes

No Comments

No Known Activations