INDEX

Explanations

references to status updates or announcements

oai_token-act-pair · gpt-3.5-turbo

references to status updates and personnel in reports

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

 borrowed

-0.74

inburgh

-0.66

jiang

-0.66

ourcing

-0.66

 Fold

-0.65

 copied

-0.65

 fold

-0.65

 rive

-0.65

 prose

-0.63

fold

-0.62

POSITIVE LOGITS

status

3.24

Status

2.40

 Status

2.19

STAT

1.16

Requirements

1.00

 STAT

0.99

Result

0.97

stat

0.97

Rating

0.94

Ready

0.89

Activations Density 0.024%

references to status updates or announcements

references to status updates and personnel in reports

No Comments

No Known Activations