INDEX

Explanations

specific formatting or layout-related terms, typically associated with web or document structure

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 4-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L4R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.4.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

insky

-0.16

 DÃŃky

-0.16

hod

-0.15

 collective

-0.15

 skip

-0.14

 Coff

-0.14

itude

-0.14

é«ĺçŃī

-0.14

ief

-0.14

 ë¹Į

-0.13

POSITIVE LOGITS

 Offensive

0.16

onia

0.15

Baz

0.15

ç©´

0.15

ÎºÎ¹

0.14

stell

0.14

Aws

0.14

ç¿

0.14

ober

0.14

apy

0.13

Activations Density 0.060%

specific formatting or layout-related terms, typically associated with web or document structure

No Comments

No Known Activations