INDEX

Explanations

mentions of Native American-related terms

oai_token-act-pair · gpt-3.5-turbo

references to Native American topics or issues

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 0-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.0.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.0.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

laden

-0.76

ammers

-0.74

zag

-0.71

ordon

-0.71

 gloom

-0.70

fax

-0.69

leck

-0.68

hall

-0.68

dam

-0.68

 Carlo

-0.68

POSITIVE LOGITS

 Native

3.85

Native

3.02

 Navajo

1.97

 Indigenous

1.88

native

1.86

 native

1.79

 aboriginal

1.69

 natives

1.66

 indigenous

1.64

 Aboriginal

1.63

Activations Density 0.022%

mentions of Native American-related terms

references to Native American topics or issues

No Comments

No Known Activations