INDEX

Explanations

words related to choices or alternatives

oai_token-act-pair · gpt-3.5-turbo

variations of the word "opt" and its related forms

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

 stamp

-0.72

 Hammer

-0.70

 Endurance

-0.68

 lasting

-0.65

 bruises

-0.64

 cracked

-0.64

 junior

-0.63

 celebrations

-0.63

 ecstasy

-0.63

 lingering

-0.62

POSITIVE LOGITS

opt

4.62

option

2.31

opted

2.19

Opt

1.87

Opt

1.51

opt

1.50

ops

1.46

options

1.38

op

1.27

opter

1.26

Activations Density 0.007%

words related to choices or alternatives

variations of the word "opt" and its related forms

No Comments

No Known Activations