INDEX

Explanations

phrases related to self-defense and use of force

oai_token-act-pair · gpt-3.5-turbo

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

raft

-0.86

 âĤ¬

-0.74

 Premium

-0.73

Spa

-0.71

 Sync

-0.71

 Transcript

-0.71

 Mold

-0.70

 Schwe

-0.69

DCS

-0.69

 Horizons

-0.68

POSITIVE LOGITS

 unarmed

1.56

 nonviolent

1.45

 violence

1.41

 combatants

1.37

 aggress

1.36

 violent

1.35

 armed

1.34

 murderers

1.33

 lethal

1.32

 gunfire

1.32

Activations Density 0.591%

phrases related to self-defense and use of force

No Comments

No Known Activations