INDEX

Explanations

URLs, particularly ones that point to Twitter

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 15-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L15R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.15.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

ustr

-0.17

ishops

-0.16

afari

-0.15

-0.14

bred

-0.14

á»ī

-0.14

.ls

-0.14

odic

-0.14

azel

-0.14

Ð»Ð°Ð³Ð°

-0.14

POSITIVE LOGITS

ı

0.15

Wr

0.14

zdy

0.14

ÄĻp

0.14

atte

0.14

 Challenger

0.14

undles

0.13

Ø®Ø¨Ø±

0.13

outu

0.13

sher

0.13

Activations Density 0.004%

URLs, particularly ones that point to Twitter

No Comments

No Known Activations