INDEX

Explanations

references to temples and places of worship

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 23-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L23R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.23.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

 Benedict

-0.15

 Lutheran

-0.15

sto

-0.14

 bible

-0.14

Ã©gor

-0.14

 ÑıÑĢ

-0.14

ael

-0.14

oo

-0.14

åº§

-0.14

DidChange

-0.13

POSITIVE LOGITS

Bah

0.18

Bah

0.18

 jenter

0.17

ogh

0.16

lessly

0.15

gbc

0.15

 Herald

0.15

 Tablets

0.15

mess

0.15

ml

0.14

Activations Density 0.003%

references to temples and places of worship

No Comments

No Known Activations