INDEX

Explanations

terms related to legal actions and law enforcement

oai_token-act-pair · gpt-3.5-turbo

phrases related to government policies and regulations

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 9-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.9.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.9.hook_resid_pre

Architecture

standard

Context Size

128

Dataset

Skylion007/openwebtext

Hook Point Layer

Activation Function

relu

Embeds

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

[|

-0.79

ruary

-0.77

 Dimensions

-0.77

optim

-0.74

terday

-0.74

 mathemat

-0.74

 Alas

-0.72

Parameter

-0.71

retched

-0.68

thank

-0.66

POSITIVE LOGITS

 lawfully

1.21

 unlawfully

1.20

 unlawful

1.19

 wrongful

1.19

 delinquent

1.14

 misdemeanor

1.13

 offender

1.09

 revocation

1.08

 criminal

1.08

 viol

1.08

Activations Density 0.629%

terms related to legal actions and law enforcement

phrases related to government policies and regulations

No Comments

No Known Activations