INDEX

Explanations

locations/events and people's titles or roles in them

oai_token-act-pair · gpt-3.5-turbo

references to time-related phrases or indicators

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 10-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.10.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.10.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

CLASS

-0.62

lda

-0.58

 thri

-0.58

dirty

-0.58

isphere

-0.56

podcast

-0.55

 bloss

-0.55

emo

-0.55

Thumbnail

-0.54

birds

-0.54

POSITIVE LOGITS

 least

1.17

 times

1.00

 midday

0.96

onement

0.90

rial

0.87

oning

0.83

 stake

0.82

kinson

0.81

oned

0.80

 halftime

0.80

Activations Density 0.096%

locations/events and people's titles or roles in them

references to time-related phrases or indicators

No Comments

No Known Activations