INDEX

Explanations

mentions or references to the word "Orange"

oai_token-act-pair · gpt-3.5-turbo

references to "Orange," likely related to a specific location or entity

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

 hurd

-0.76

irm

-0.70

--

-0.66

dj

-0.65

sha

-0.65

 preempt

-0.64

 streng

-0.64

lihood

-0.63

)--

-0.63

hw

-0.63

POSITIVE LOGITS

 Orange

3.94

Orange

3.37

orange

2.03

 orange

1.92

 oranges

1.59

 Purple

1.49

 Irvine

1.41

 Anaheim

1.41

 Yellow

1.35

 Riverside

1.33

Activations Density 0.027%

mentions or references to the word "Orange"

references to "Orange," likely related to a specific location or entity

No Comments

No Known Activations