INDEX

Explanations

words related to legal violations or breaches of law

oai_token-act-pair · gpt-3.5-turbo

references to legal violations or breaches of rules and regulations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

tal

-0.80

stars

-0.71

pher

-0.69

iences

-0.69

stra

-0.68

pine

-0.67

 confidently

-0.65

omics

-0.65

landers

-0.64

 aged

-0.64

POSITIVE LOGITS

 violation

3.76

 violations

2.29

viol

2.13

 violating

2.02

 breach

1.94

 violate

1.84

 viol

1.78

 violated

1.76

 Viol

1.72

 violates

1.66

Activations Density 0.013%

words related to legal violations or breaches of law

references to legal violations or breaches of rules and regulations

No Comments

No Known Activations