INDEX

Explanations

language related to swearing and racial slurs

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

orgh

-0.17

hare

-0.14

jezd

-0.14

ÑģÑĤÐ¾ÑĢ

-0.13

yre

-0.13

æĭ

-0.13

 yearly

-0.13

 Wilkinson

-0.13

qq

-0.13

POSITIVE LOGITS

 prof

0.49

 swear

0.46

 curse

0.44

 swearing

0.43

 curs

0.40

 curses

0.36

 Prof

0.36

prof

0.35

 obsc

0.34

 Curse

0.34

Activations Density 0.164%

language related to swearing and racial slurs

No Comments

No Known Activations