INDEX

Explanations

terms related to violent actions and attacks on people or property

oai_token-act-pair · gpt-3.5-turbo

references to violence or attacks against individuals, particularly police and civilians

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 renaissance

-0.72

 Conj

-0.66

 Farming

-0.65

joining

-0.64

lopp

-0.62

join

-0.62

 reinvent

-0.62

imester

-0.59

ISTORY

-0.58

okin

-0.57

POSITIVE LOGITS

 targets

0.88

 unarmed

0.86

itals

0.79

 helpless

0.78

 innocent

0.74

 senseless

0.73

soever

0.72

 unconscious

0.72

 innoc

0.70

 indiscrim

0.69

Activations Density 0.769%

terms related to violent actions and attacks on people or property

references to violence or attacks against individuals, particularly police and civilians

No Comments

No Known Activations