INDEX

Explanations

sentences ending with a full stop

oai_token-act-pair · gpt-3.5-turbo

sentence-like structures, particularly those that suggest a statement or assertion

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

 distinguished

-0.79

 activ

-0.74

 scattered

-0.74

 classified

-0.73

 effected

-0.73

 eager

-0.72

 purposes

-0.72

 broadly

-0.70

 observers

-0.70

uly

-0.70

POSITIVE LOGITS

 Beware

1.35

 Conclusion

1.22

[+

1.15

 Avoid

1.09

Why

1.09

 Deter

1.05

 Provide

1.05

 Thou

1.05

 Increase

1.02

 Explain

1.02

Activations Density 0.062%

sentences ending with a full stop

sentence-like structures, particularly those that suggest a statement or assertion

No Comments

No Known Activations