INDEX

Explanations

people's names

oai_token-act-pair · gpt-3.5-turbo

repeated mentions of the name "Jo."

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

 Spectrum

-0.80

 disclosure

-0.77

 disclosures

-0.69

 vectors

-0.69

 Reset

-0.64

 carriers

-0.64

 vector

-0.63

 inaccurate

-0.63

 microsc

-0.62

 Medium

-0.60

POSITIVE LOGITS

jo

4.50

JO

2.15

Jo

1.76

ja

1.62

jas

1.57

joining

1.48

je

1.40

ji

1.39

jon

1.34

1.33

Activations Density 0.010%

people's names

repeated mentions of the name "Jo."

No Comments

No Known Activations