INDEX

Explanations

instances of events or instances that validate or confirm a specific outcome or situation

oai_token-act-pair · gpt-3.5-turbo

instances of the word "prove" and its various forms, indicating validation or confirmation of claims

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

IFrame

Link

Not in Any Lists

No Comments

Negative Logits

letal

-0.80

umbn

-0.73

orget

-0.73

pta

-0.73

ulkan

-0.71

arta

-0.69

VIDIA

-0.68

lished

-0.67

enaries

-0.64

etting

-0.64

POSITIVE LOGITS

 invaluable

1.20

 resilient

1.18

 irresistible

1.15

 elusive

1.10

 fruitful

1.09

 ineffective

1.08

 prophetic

0.99

 incapable

0.99

 useful

0.99

 adept

0.98

Activations Density 0.043%

instances of events or instances that validate or confirm a specific outcome or situation

instances of the word "prove" and its various forms, indicating validation or confirmation of claims

No Comments

No Known Activations