INDEX

Explanations

phrases related to government policies or actions

oai_token-act-pair · gpt-3.5-turbo

structured data or report-like components

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 6-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.6.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.6.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

 lifes

-0.65

 receiving

-0.60

 spitting

-0.59

 intended

-0.59

 legally

-0.59

 diving

-0.59

 inhal

-0.59

 ticket

-0.58

 treasury

-0.57

 footing

-0.57

POSITIVE LOGITS

Conclusion

1.46

Finally

1.44

Lastly

1.43

CONCLUS

1.23

Another

1.14

Anyway

1.10

Overall

1.10

Meanwhile

1.10

Still

1.07

 Lastly

1.06

Activations Density 0.696%

phrases related to government policies or actions

structured data or report-like components

No Comments

No Known Activations