INDEX

Explanations

distances or measurements

oai_token-act-pair · gpt-3.5-turbo

references to physical distances

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

ito

-0.69

hern

-0.67

 Audit

-0.66

gob

-0.66

Ol

-0.64

enh

-0.64

 authored

-0.63

 Nich

-0.62

apps

-0.62

payers

-0.62

POSITIVE LOGITS

 distance

4.17

 distances

3.08

 Distance

2.79

distance

2.59

Distance

2.34

 proximity

1.48

 radius

1.43

 altitude

1.36

 distant

1.29

 diameter

1.21

Activations Density 0.013%

distances or measurements

references to physical distances

No Comments

No Known Activations