INDEX

Explanations

terms related to compliance and security in healthcare contexts

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

å³°

-0.15

WWW

-0.14

Insets

-0.13

 nÃło

-0.13

æĹ¦

-0.13

.lu

-0.13

,readonly

-0.13

/Dk

-0.13

 afflicted

-0.13

.spy

-0.13

POSITIVE LOGITS

 state

0.54

state

0.42

 advanced

0.41

 cutting

0.38

 leading

0.38

-state

0.36

(state

0.35

.state

0.35

 State

0.34

advanced

0.32

Activations Density 0.556%

terms related to compliance and security in healthcare contexts

No Comments

No Known Activations