INDEX

Explanations

phrases and terms emphasizing priority and safety in various contexts

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

ssi

-0.16

izin

-0.16

 Prelude

-0.15

ignet

-0.14

 Helpful

-0.14

aklÄ±

-0.14

Pad

-0.14

achel

-0.14

ropol

-0.14

.Provider

-0.14

POSITIVE LOGITS

 priority

0.73

priority

0.61

 priorities

0.60

 Priority

0.59

Priority

0.57

_priority

0.50

 priorit

0.49

.priority

0.48

 prioritize

0.47

top

0.45

Activations Density 0.147%

phrases and terms emphasizing priority and safety in various contexts

No Comments

No Known Activations