INDEX

Explanations

terms related to hate and hate crimes

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With LLAMA3.1-8B @ 26-llamascope-res-32k

Configuration

fnlp/Llama3_1-8B-Base-LXR-8x/Llama3_1-8B-Base-L26R-8x

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

cerebras/SlimPajama-627B

Features

32,768

Data Type

bfloat16

Hook Name

blocks.26.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

éĥİ

-0.16

erie

-0.15

ettle

-0.14

.scalablytyped

-0.14

erable

-0.14

 Ð·Ð°ÑģÑĤ

-0.14

tes

-0.14

Ð´Ð¸Ð½Ð°

-0.14

iere

-0.14

ians

-0.13

POSITIVE LOGITS

 speech

0.35

fully

0.31

 Speech

0.27

 crime

0.26

peech

0.26

fulness

0.25

filled

0.25

 crimes

0.25

Speech

0.24

speech

0.24

Activations Density 0.008%

terms related to hate and hate crimes

No Comments

No Known Activations