INDEX

Explanations

mentions of people or characters

oai_token-act-pair · gpt-3.5-turbo

mentions of witnesses and individuals involved in incidents or events

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 11-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.11.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.11.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

 actionGroup

-0.79

Welcome

-0.72

venge

-0.67

 reforming

-0.62

 Welcome

-0.61

Effective

-0.60

needs

-0.60

 dear

-0.59

Frameworks

-0.58

 Flavoring

-0.58

POSITIVE LOGITS

 claimed

1.49

 stated

1.47

 alleges

1.43

 allege

1.37

 claims

1.29

 alleged

1.29

 recounted

1.26

 speculated

1.25

 said

1.22

 asserted

1.20

Activations Density 0.398%

mentions of people or characters

mentions of witnesses and individuals involved in incidents or events

No Comments

No Known Activations