INDEX

Explanations

proper nouns, specifically names of places and entities

oai_token-act-pair · gpt-3.5-turbo

proper nouns related to places, political parties, and significant individuals

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

New Auto-Interp

Top Features by Cosine Similarity

Comparing With GPT2-SMALL @ 5-res-jb

Configuration

jbloom/GPT2-Small-SAEs-Reformatted/blocks.5.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.5.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

 misdemeanor

-0.91

 NAACP

-0.85

 statewide

-0.83

 Boone

-0.82

 nonprofits

-0.81

 Wyoming

-0.81

Iowa

-0.79

 felony

-0.79

 feds

-0.78

 Spokane

-0.77

POSITIVE LOGITS

alys

1.20

 analys

1.03

 recognise

1.03

avour

1.02

 analyse

0.98

 Whilst

0.97

 mould

0.94

 organisers

0.94

 organise

0.93

 metre

0.92

Activations Density 0.825%

proper nouns, specifically names of places and entities

proper nouns related to places, political parties, and significant individuals

No Comments

No Known Activations