INDEX

Explanations

references to self-harm and suicide

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

exus

-0.17

 Sponsored

-0.14

.Sdk

-0.14

acle

-0.14

Deg

-0.14

sponsor

-0.14

ACLE

-0.14

.learn

-0.14

olver

-0.14

acro

-0.14

POSITIVE LOGITS

 suicide

0.60

 Suicide

0.52

-su

0.52

su

0.50

Su

0.50

 suicides

0.49

 commit

0.49

_su

0.49

Su

0.48

 suic

0.45

Activations Density 0.099%

references to self-harm and suicide

No Comments

No Known Activations