INDEX

Explanations

terms related to legal or official documents

oai_token-act-pair · gpt-3.5-turbo

contextually relevant phrases or clauses

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 8-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.8.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.8.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

•Partition Features Layer 8 GPT2 Small

" token" vs "Token"

No Comments

Negative Logits

 ende

-0.85

 neighb

-0.80

 quir

-0.80

 tram

-0.79

 clin

-0.76

 derby

-0.76

 emancipation

-0.75

 lifes

-0.74

 relegation

-0.74

 escaping

-0.74

POSITIVE LOGITS

Quote

1.48

https

1.48

http

1.47

âĢ¢

1.36

If

1.34

"âĢ¦

1.33

From

1.33

When

1.33

We

1.32

Dear

1.32

Activations Density 0.119%

terms related to legal or official documents

contextually relevant phrases or clauses

No Comments

No Known Activations