INDEX

Explanations

phrases related to significant actions or events

oai_token-act-pair · gpt-3.5-turbo

phrases related to acts of wrongdoing or violence

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

 Flavoring

-0.81

ials

-0.75

 sshd

-0.73

 corners

-0.69

 Scores

-0.68

 Centers

-0.66

 ceilings

-0.66

nces

-0.66

bid

-0.66

 Extras

-0.64

POSITIVE LOGITS

 kindness

1.16

 sabotage

1.14

 heroism

1.09

 vandalism

1.04

 aggression

1.02

 defiance

0.97

 bravery

0.96

 violence

0.96

 desperation

0.95

 generosity

0.90

Activations Density 0.074%

phrases related to significant actions or events

phrases related to acts of wrongdoing or violence

No Comments

No Known Activations