INDEX

Explanations

proper names, specifically those related to Gordon

oai_token-act-pair · gpt-3.5-turbo

references to the name "Gordon," particularly in relation to notable figures or events

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

ace

-0.70

NCT

-0.65

pus

-0.64

 migrate

-0.62

 rider

-0.61

cap

-0.61

 airplanes

-0.61

 Authorization

-0.61

xual

-0.61

 enchant

-0.60

POSITIVE LOGITS

 Gordon

3.75

Gordon

2.94

 Ramsay

1.41

 Freeman

1.39

 Cameron

1.37

 Blair

1.35

 Dudley

1.22

 Boyd

1.21

 Bradford

1.20

 Campbell

1.18

Activations Density 0.012%

proper names, specifically those related to Gordon

references to the name "Gordon," particularly in relation to notable figures or events

No Comments

No Known Activations