INDEX

Explanations

references to individuals, particularly those with honorifics or titles

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

umper

-0.15

ãĥ¼ãĥį

-0.15

Éµ

-0.15

tas

-0.15

obble

-0.15

orris

-0.15

ÐµÐ»ÐµÑĦ

-0.14

asca

-0.14

utenberg

-0.14

oodoo

-0.14

POSITIVE LOGITS

0.15

 redraw

0.15

anna

0.15

0.14

empt

0.13

 Authors

0.13

 Laure

0.13

riot

0.13

Activations Density 0.035%

references to individuals, particularly those with honorifics or titles

No Comments

No Known Activations