INDEX

Explanations

verbs in infinitive form following particular pronouns

oai_token-act-pair · gpt-3.5-turbo

phrases indicating actions and intentions

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

 civilizations

-0.96

colored

-0.90

ADVERTISEMENT

-0.86

izes

-0.85

 behaviors

-0.84

defense

-0.83

imize

-0.83

 characterization

-0.83

 offenses

-0.80

 offense

-0.79

POSITIVE LOGITS

 criticised

1.65

 Â£

1.47

 emphas

1.43

 recognised

1.42

 learnt

1.40

 apologised

1.39

 analysed

1.32

 scept

1.31

 recognise

1.29

 honoured

1.29

Activations Density 0.395%

verbs in infinitive form following particular pronouns

phrases indicating actions and intentions

No Comments

No Known Activations