INDEX

Explanations

references to specific quantities or measurements, particularly doses of something

oai_token-act-pair · gpt-3.5-turbo

references to "dose" or varying quantities related to medication or substances

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

Nou

-0.67

Pal

-0.66

backs

-0.65

boards

-0.65

wal

-0.63

LCS

-0.63

 Strongh

-0.62

behind

-0.62

hips

-0.61

ans

-0.61

POSITIVE LOGITS

 dose

3.76

 doses

3.06

 dosage

2.50

dose

2.48

dos

1.61

 horm

1.27

 overdose

1.26

 injection

1.24

 injections

1.19

 potency

1.18

Activations Density 0.015%

references to specific quantities or measurements, particularly doses of something

references to "dose" or varying quantities related to medication or substances

No Comments

No Known Activations