INDEX

Explanations

phrases that indicate polling data and statistical results

oai_token-act-pair · gpt-4o-mini Triggered by @bot

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res_fs24576-jb

Configuration

jbloom/GPT2-Small-Feature-Splitting-Experiment-Layer-8/blocks.8.hook_resid_pre_24576

Prompts (Dashboard)

24,576 prompts, 128 tokens each

Dataset (Dashboard)

Skylion007/openwebtext

Features

24,576

Data Type

torch.float32

Hook Point

blocks.8.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

 downright

-0.82

 swat

-0.81

 unstoppable

-0.79

 shove

-0.78

 stubborn

-0.78

 choking

-0.78

 vanishing

-0.78

 popping

-0.77

 futile

-0.76

 calming

-0.76

POSITIVE LOGITS

 Additionally

1.23

<|endoftext|>

1.20

 Afterwards

1.15

 Additional

1.13

 Each

1.12

Its

1.06

 Approximately

1.06

 Dates

1.05

 During

1.05

 Upon

1.03

Activations Density 0.392%

phrases that indicate polling data and statistical results

No Comments

No Known Activations