INDEX

Explanations

mentions of criminal activities or incidents

oai_token-act-pair · gpt-3.5-turbo

instances of numerical data and timestamps in relation to events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 scept

-0.91

 favour

-0.84

 censorship

-0.81

 censor

-0.81

 princip

-0.77

 endorsing

-0.74

 monet

-0.74

 publishing

-0.74

 negotiating

-0.73

 publish

-0.73

POSITIVE LOGITS

Police

1.20

Investigators

1.18

Authorities

1.12

Officers

1.09

RELATED

1.03

According

1.02

WARNING

1.02

However

1.01

Officials

1.01

ccording

1.00

Activations Density 0.578%

mentions of criminal activities or incidents

instances of numerical data and timestamps in relation to events

No Comments

No Known Activations