INDEX

Explanations

words related to legal entities or locations

oai_token-act-pair · gpt-3.5-turbo

references to specific individuals and titles in the context of governance or authority

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 1-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.1.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.1.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

 Donation

-0.78

ahime

-0.77

etter

-0.75

 Catalyst

-0.73

Hardware

-0.70

IPM

-0.67

[|

-0.67

eri

-0.65

Connector

-0.64

 Investor

-0.64

POSITIVE LOGITS

Dum

1.10

pty

0.83

rend

0.69

 Pret

0.66

oeuv

0.66

amina

0.66

 chau

0.65

uple

0.65

arg

0.65

 imperson

0.63

Activations Density 0.020%

words related to legal entities or locations

references to specific individuals and titles in the context of governance or authority

No Comments

No Known Activations