INDEX

Explanations

phrases related to destruction, violence, and threats directed towards others

oai_token-act-pair · gpt-3.5-turbo

violent actions or threats

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.hook_resid_pre

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.10.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

eret

-0.74

 Advis

-0.68

":["

-0.67

iola

-0.63

 Advice

-0.63

 heightened

-0.63

iera

-0.62

soDeliveryDate

-0.62

CES

-0.61

LH

-0.61

POSITIVE LOGITS

 alive

1.24

 senseless

1.06

 limb

1.03

 utterly

0.99

 altogether

0.98

 whole

0.96

 merciless

0.92

 Alive

0.90

 entirely

0.90

 inciner

0.89

Activations Density 0.283%

phrases related to destruction, violence, and threats directed towards others

violent actions or threats

No Comments

No Known Activations