INDEX

Explanations

terms related to things that are hidden, undisclosed, or not widely known

oai_token-act-pair · gpt-3.5-turbo

prefixes related to negation or lack

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

PlotsExplanationShow Test FieldDefault Test Text

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

=-=-=-=-=-=-=-=-

-0.75

 bull

-0.74

 bulls

-0.72

 sarc

-0.67

 sucking

-0.67

 simulator

-0.63

 initials

-0.62

 Knights

-0.62

 straw

-0.61

 rooting

-0.60

POSITIVE LOGITS

leased

1.58

achable

1.48

ported

1.47

ason

1.30

ached

1.26

peat

1.25

vised

1.20

ired

1.13

acted

1.10

emed

1.03

Activations Density 0.028%

terms related to things that are hidden, undisclosed, or not widely known

prefixes related to negation or lack

No Comments

No Known Activations