INDEX

Explanations

numbers and proper nouns

oai_token-act-pair · gpt-3.5-turbo

instances of strong emotions or dramatic situations

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

 complementary

-0.82

 immigration

-0.80

 fallout

-0.79

 licensing

-0.77

 specifically

-0.77

 alleged

-0.77

 Reuters

-0.77

cod

-0.75

 reportedly

-0.75

 similar

-0.75

POSITIVE LOGITS

Cause

1.70

And

1.58

Where

1.47

Who

1.45

Tell

1.43

Which

1.41

Everybody

1.39

Come

1.39

Unt

1.39

Give

1.38

Activations Density 0.133%

numbers and proper nouns

instances of strong emotions or dramatic situations

No Comments

No Known Activations